Trying to add spacer at the end of the transcript

docker-compose.example.yml

@@ -12,6 +12,9 @@ services:
       OPENAI_API_KEY: ${OPENAI_API_KEY:-}
       ANTHROPIC_API_KEY: ${ANTHROPIC_API_KEY:-}
       XAI_API_KEY: ${XAI_API_KEY:-}
+      HERMES_AGENT_API_BASE_URL: ${HERMES_AGENT_API_BASE_URL:-http://127.0.0.1:8642/v1}
+      HERMES_AGENT_API_KEY: ${HERMES_AGENT_API_KEY:-}
+      HERMES_AGENT_MODEL: ${HERMES_AGENT_MODEL:-}
       EXA_API_KEY: ${EXA_API_KEY:-}
       CHAT_WEB_SEARCH_ENGINE: ${CHAT_WEB_SEARCH_ENGINE:-exa}
       SEARXNG_BASE_URL: ${SEARXNG_BASE_URL:-}

docs/api/rest.md

@@ -33,11 +33,13 @@ Chat upload limits:
   "providers": {
     "openai": { "models": ["gpt-4.1-mini"], "loadedAt": "2026-02-14T00:00:00.000Z", "error": null },
     "anthropic": { "models": ["claude-3-5-sonnet-latest"], "loadedAt": null, "error": null },
-    "xai": { "models": ["grok-3-mini"], "loadedAt": null, "error": null }
+    "xai": { "models": ["grok-3-mini"], "loadedAt": null, "error": null },
+    "hermes-agent": { "models": ["hermes-agent"], "loadedAt": null, "error": null }
   }
 }
 ```
 - OpenAI model lists are filtered to models that are expected to work with the backend's Responses API implementation.
+- `hermes-agent` is included only when `HERMES_AGENT_API_KEY` is configured. Set it to Hermes `API_SERVER_KEY`, or any non-empty value if that local server does not require auth. `HERMES_AGENT_API_BASE_URL` defaults to `http://127.0.0.1:8642/v1`; set `HERMES_AGENT_MODEL` only when you need an additional fallback/override model id.
 
 ## Active Runs
 
@@ -65,7 +67,7 @@ Behavior notes:
 ```json
 {
   "title": "optional title",
-  "provider": "optional openai|anthropic|xai",
+  "provider": "optional openai|anthropic|xai|hermes-agent",
   "model": "optional model id",
   "messages": [
     {
@@ -152,7 +154,7 @@ Notes:
 ```json
 {
   "chatId": "optional-chat-id",
-  "provider": "openai|anthropic|xai",
+  "provider": "openai|anthropic|xai|hermes-agent",
   "model": "string",
   "messages": [
     {
@@ -206,11 +208,12 @@ Behavior notes:
 - Text files are forwarded as explicit text blocks rather than provider-managed file references. Large text attachments should already be truncated client-side before submission.
 - For `openai`, backend calls OpenAI's Responses API and enables internal tool use with an internal system instruction.
 - For `xai`, backend calls xAI's OpenAI-compatible Chat Completions API and enables internal tool use with the same internal system instruction.
+- For `hermes-agent`, backend calls the configured Hermes Agent OpenAI-compatible Chat Completions API without adding Sybil-managed tool definitions; Hermes Agent handles its own tools server-side.
 - For `openai`, image attachments are sent as Responses `input_image` items and text attachments are sent as `input_text` items.
-- For `xai`, image attachments are sent as Chat Completions content parts alongside text.
+- For `xai` and `hermes-agent`, image attachments are sent as Chat Completions content parts alongside text.
 - For `openai`, Responses calls that can enter the server-managed tool loop use `store: true` so reasoning and function-call items can be passed between tool rounds.
 - For `anthropic`, image attachments are sent as Messages API `image` blocks using base64 source data; text attachments are added as `text` blocks.
-- Available tool calls for chat: `web_search` and `fetch_url`. When `CHAT_CODEX_TOOL_ENABLED=true`, `codex_exec` is also available. When `CHAT_SHELL_TOOL_ENABLED=true`, `shell_exec` is also available.
+- Available Sybil-managed tool calls for `openai` and `xai`: `web_search` and `fetch_url`. When `CHAT_CODEX_TOOL_ENABLED=true`, `codex_exec` is also available. When `CHAT_SHELL_TOOL_ENABLED=true`, `shell_exec` is also available.
 - `web_search` returns ranked results with per-result summaries/snippets. Its backend engine is selected by `CHAT_WEB_SEARCH_ENGINE` (`exa` default, or `searxng` with `SEARXNG_BASE_URL` set). SearXNG mode requires the instance to allow `format=json`.
 - `fetch_url` fetches a URL and returns plaintext page content (HTML converted to text server-side).
 - `codex_exec` delegates coding, shell, repository inspection, and other complex software tasks to a persistent remote Codex CLI workspace over SSH. The server runs `codex exec --dangerously-bypass-approvals-and-sandbox --skip-git-repo-check <non-interactive wrapped prompt>` on the configured devbox inside `CHAT_CODEX_REMOTE_WORKDIR`, with SSH stdin closed.
@@ -311,9 +314,9 @@ Behavior notes:
   "title": null,
   "createdAt": "...",
   "updatedAt": "...",
-  "initiatedProvider": "openai|anthropic|xai|null",
+  "initiatedProvider": "openai|anthropic|xai|hermes-agent|null",
   "initiatedModel": "string|null",
-  "lastUsedProvider": "openai|anthropic|xai|null",
+  "lastUsedProvider": "openai|anthropic|xai|hermes-agent|null",
   "lastUsedModel": "string|null"
 }
 ```
@@ -359,9 +362,9 @@ Behavior notes:
   "title": null,
   "createdAt": "...",
   "updatedAt": "...",
-  "initiatedProvider": "openai|anthropic|xai|null",
+  "initiatedProvider": "openai|anthropic|xai|hermes-agent|null",
   "initiatedModel": "string|null",
-  "lastUsedProvider": "openai|anthropic|xai|null",
+  "lastUsedProvider": "openai|anthropic|xai|hermes-agent|null",
   "lastUsedModel": "string|null",
   "messages": [Message]
 }

docs/api/streaming-chat.md

@@ -21,7 +21,7 @@ Authentication:
 {
   "chatId": "optional-chat-id",
   "persist": true,
-  "provider": "openai|anthropic|xai",
+  "provider": "openai|anthropic|xai|hermes-agent",
   "model": "string",
   "messages": [
     {
@@ -152,8 +152,9 @@ For `persist: false` streams, `chatId` and `callId` are `null`.
 
 - `openai`: backend uses OpenAI's Responses API and may execute internal function tool calls (`web_search`, `fetch_url`, optional `codex_exec`, and optional `shell_exec`) before producing final text.
 - `xai`: backend uses xAI's OpenAI-compatible Chat Completions API and may execute the same internal tool calls before producing final text.
+- `hermes-agent`: backend uses the configured Hermes Agent OpenAI-compatible Chat Completions API. Sybil does not add its own tool definitions for this provider; Hermes Agent handles its own tools server-side. Custom Hermes stream events are normalized away unless they produce text deltas in this SSE contract.
 - `openai`: image attachments are sent as Responses `input_image` items; text attachments are sent as `input_text` items.
-- `xai`: image attachments are sent as Chat Completions content parts; text attachments are inlined as text parts.
+- `xai` and `hermes-agent`: image attachments are sent as Chat Completions content parts; text attachments are inlined as text parts.
 - `openai`: Responses calls that can enter the server-managed tool loop use `store: true` so reasoning and function-call items can be passed between tool rounds.
 - `anthropic`: streamed via event stream; emits `delta` from `content_block_delta` with `text_delta`. Image attachments are sent as base64 `image` blocks and text attachments are appended as `text` blocks.
 - `web_search` uses `CHAT_WEB_SEARCH_ENGINE` (`exa` default, or `searxng` with `SEARXNG_BASE_URL` set). SearXNG mode requires the instance to allow `format=json`. This only affects chat-mode tool calls, not search-mode endpoints.

ios/AGENTS.md

@@ -51,3 +51,4 @@ Instructions for work under `/Users/buzzert/src/sybil-2/ios`.
   - OpenAI: `gpt-4.1-mini`
   - Anthropic: `claude-3-5-sonnet-latest`
   - xAI: `grok-3-mini`
+  - Hermes Agent: `hermes-agent`

ios/Apps/Sybil/Info.plist

@@ -0,0 +1,17 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<!DOCTYPE plist PUBLIC "-//Apple//DTD PLIST 1.0//EN" "http://www.apple.com/DTDs/PropertyList-1.0.dtd">
+<plist version="1.0">
+<dict>
+	<key>UIApplicationShortcutItems</key>
+	<array>
+		<dict>
+			<key>UIApplicationShortcutItemType</key>
+			<string>net.buzzert.sybil2.quick-question</string>
+			<key>UIApplicationShortcutItemTitle</key>
+			<string>Quick question</string>
+            <key>UIApplicationShortcutItemIconSymbolName</key>
+            <string>sparkles</string>
+		</dict>
+	</array>
+</dict>
+</plist>

ios/Apps/Sybil/Sources/SybilApp.swift

@@ -5,6 +5,8 @@ import UIKit
 @main
 struct SybilApp: App
 {
+    @UIApplicationDelegateAdaptor(SybilAppDelegate.self) private var appDelegate
+
     var body: some Scene {
         WindowGroup {
             SplitView()
@@ -14,3 +16,79 @@ struct SybilApp: App
         }
     }
 }
+
+@MainActor
+final class SybilAppDelegate: NSObject, UIApplicationDelegate {
+    func application(
+        _ application: UIApplication,
+        didFinishLaunchingWithOptions launchOptions: [UIApplication.LaunchOptionsKey: Any]? = nil
+    ) -> Bool {
+        SybilHomeScreenQuickActionHandler.configureQuickActions()
+        return true
+    }
+
+    func application(
+        _ application: UIApplication,
+        configurationForConnecting connectingSceneSession: UISceneSession,
+        options: UIScene.ConnectionOptions
+    ) -> UISceneConfiguration {
+        let configuration = UISceneConfiguration(
+            name: "Default Configuration",
+            sessionRole: connectingSceneSession.role
+        )
+        configuration.delegateClass = SybilSceneDelegate.self
+        return configuration
+    }
+
+    func application(
+        _ application: UIApplication,
+        performActionFor shortcutItem: UIApplicationShortcutItem,
+        completionHandler: @escaping (Bool) -> Void
+    ) {
+        completionHandler(SybilHomeScreenQuickActionHandler.handle(shortcutItem))
+    }
+}
+
+@MainActor
+final class SybilSceneDelegate: NSObject, UIWindowSceneDelegate {
+    func scene(
+        _ scene: UIScene,
+        willConnectTo session: UISceneSession,
+        options connectionOptions: UIScene.ConnectionOptions
+    ) {
+        if let shortcutItem = connectionOptions.shortcutItem {
+            _ = SybilHomeScreenQuickActionHandler.handle(shortcutItem)
+        }
+    }
+
+    func windowScene(
+        _ windowScene: UIWindowScene,
+        performActionFor shortcutItem: UIApplicationShortcutItem,
+        completionHandler: @escaping (Bool) -> Void
+    ) {
+        completionHandler(SybilHomeScreenQuickActionHandler.handle(shortcutItem))
+    }
+
+    func sceneWillResignActive(_ scene: UIScene) {
+        SybilHomeScreenQuickActionHandler.configureQuickActions()
+    }
+}
+
+@MainActor
+private enum SybilHomeScreenQuickActionHandler {
+    static func configureQuickActions() {
+        // The quick question action is static in Info.plist so it is available before first launch.
+        UIApplication.shared.shortcutItems = []
+    }
+
+    static func handle(_ shortcutItem: UIApplicationShortcutItem) -> Bool {
+        guard shortcutItem.type == SybilHomeScreenQuickAction.quickQuestionType else {
+            return false
+        }
+
+        Task { @MainActor in
+            SybilQuickActionRouter.shared.requestQuickQuestionPresentation()
+        }
+        return true
+    }
+}

ios/Apps/Sybil/project.yml

@@ -22,9 +22,10 @@ targets:
         SUPPORTS_MAC_DESIGNED_FOR_IPHONE_IPAD: NO
         TARGETED_DEVICE_FAMILY: "1,2,6"
         GENERATE_INFOPLIST_FILE: YES
+        INFOPLIST_FILE: Apps/Sybil/Info.plist
         ASSETCATALOG_COMPILER_APPICON_NAME: AppIcon
-        MARKETING_VERSION: 1.5
+        MARKETING_VERSION: 1.7
-        CURRENT_PROJECT_VERSION: 6
+        CURRENT_PROJECT_VERSION: 8
         INFOPLIST_KEY_CFBundleDisplayName: Sybil
         INFOPLIST_KEY_ITSAppUsesNonExemptEncryption: NO
         INFOPLIST_KEY_UIApplicationSupportsIndirectInputEvents: YES

ios/Packages/Sybil/Sources/Sybil/SplitView.swift

@@ -2,10 +2,14 @@ import SwiftUI
 
 public struct SplitView: View {
     @State private var viewModel = SybilViewModel()
+    @ObservedObject private var quickActionRouter = SybilQuickActionRouter.shared
     @Environment(\.horizontalSizeClass) private var horizontalSizeClass
     @Environment(\.scenePhase) private var scenePhase
     @State private var shouldRefreshOnForeground = false
     @State private var composerFocusRequest = 0
+    @State private var quickQuestionFocusRequest = 0
+    @State private var hasPendingQuickQuestionPresentation = false
+    @State private var isQuickQuestionPresented = false
     @State private var columnVisibility: NavigationSplitViewVisibility = .automatic
 
     private var keyboardActions: SybilKeyboardActions? {
@@ -74,8 +78,28 @@ public struct SplitView: View {
         .font(.sybil(.body))
         .preferredColorScheme(.dark)
         .focusedSceneValue(\.sybilKeyboardActions, keyboardActions)
+        .sheet(isPresented: $isQuickQuestionPresented, onDismiss: handleQuickQuestionDismissed) {
+            SybilQuickQuestionView(
+                viewModel: viewModel,
+                focusRequest: quickQuestionFocusRequest
+            )
+            .presentationDragIndicator(.visible)
+        }
         .task {
             await viewModel.bootstrap()
+            presentPendingQuickQuestionIfPossible()
+        }
+        .onReceive(quickActionRouter.$quickQuestionPresentationRequest) { request in
+            guard request > 0 else {
+                return
+            }
+            queueQuickQuestionPresentation()
+        }
+        .onChange(of: viewModel.isCheckingSession) { _, _ in
+            presentPendingQuickQuestionIfPossible()
+        }
+        .onChange(of: viewModel.isAuthenticated) { _, _ in
+            presentPendingQuickQuestionIfPossible()
         }
         .onChange(of: scenePhase) { _, nextPhase in
             switch nextPhase {
@@ -112,6 +136,28 @@ public struct SplitView: View {
             columnVisibility = .all
         }
     }
+
+    private func queueQuickQuestionPresentation() {
+        hasPendingQuickQuestionPresentation = true
+        presentPendingQuickQuestionIfPossible()
+    }
+
+    private func presentPendingQuickQuestionIfPossible() {
+        guard hasPendingQuickQuestionPresentation,
+              !viewModel.isCheckingSession,
+              viewModel.isAuthenticated
+        else {
+            return
+        }
+
+        hasPendingQuickQuestionPresentation = false
+        quickQuestionFocusRequest += 1
+        isQuickQuestionPresented = true
+    }
+
+    private func handleQuickQuestionDismissed() {
+        viewModel.cancelQuickQuestion()
+    }
 }
 
 public struct SybilCommands: Commands {

ios/Packages/Sybil/Sources/Sybil/SybilAPIClient.swift

@@ -49,11 +49,16 @@ actor SybilAPIClient: SybilAPIClienting {
         return response.chats
     }
 
-    func createChat(title: String? = nil) async throws -> ChatSummary {
+    func createChat(
+        title: String? = nil,
+        provider: Provider? = nil,
+        model: String? = nil,
+        messages: [CompletionRequestMessage]? = nil
+    ) async throws -> ChatSummary {
         let response = try await request(
             "/v1/chats",
             method: "POST",
-            body: AnyEncodable(ChatCreateBody(title: title)),
+            body: AnyEncodable(ChatCreateBody(title: title, provider: provider, model: model, messages: messages)),
             responseType: ChatCreateResponse.self
         )
         return response.chat
@@ -617,6 +622,7 @@ actor SybilAPIClient: SybilAPIClienting {
 
 struct CompletionStreamRequest: Codable, Sendable {
     var chatId: String?
+    var persist: Bool? = nil
     var provider: Provider
     var model: String
     var messages: [CompletionRequestMessage]
@@ -624,6 +630,9 @@ struct CompletionStreamRequest: Codable, Sendable {
 
 private struct ChatCreateBody: Encodable {
     var title: String?
+    var provider: Provider?
+    var model: String?
+    var messages: [CompletionRequestMessage]?
 }
 
 private struct SearchCreateBody: Encodable {

ios/Packages/Sybil/Sources/Sybil/SybilAPIClienting.swift

@@ -3,7 +3,12 @@ import Foundation
 protocol SybilAPIClienting: Sendable {
     func verifySession() async throws -> AuthSession
     func listChats() async throws -> [ChatSummary]
-    func createChat(title: String?) async throws -> ChatSummary
+    func createChat(
+        title: String?,
+        provider: Provider?,
+        model: String?,
+        messages: [CompletionRequestMessage]?
+    ) async throws -> ChatSummary
     func getChat(chatID: String) async throws -> ChatDetail
     func deleteChat(chatID: String) async throws
     func suggestChatTitle(chatID: String, content: String) async throws -> ChatSummary
@@ -32,3 +37,9 @@ protocol SybilAPIClienting: Sendable {
         onEvent: @escaping @Sendable (SearchStreamEvent) async -> Void
     ) async throws
 }
+
+extension SybilAPIClienting {
+    func createChat(title: String?) async throws -> ChatSummary {
+        try await createChat(title: title, provider: nil, model: nil, messages: nil)
+    }
+}

ios/Packages/Sybil/Sources/Sybil/SybilChatTranscriptView.swift

@@ -7,6 +7,9 @@ struct SybilChatTranscriptView: View {
     var isSending: Bool
     var topContentInset: CGFloat = 0
     var bottomContentInset: CGFloat = 0
+    var tailSpacerHeight: CGFloat = 0
+    var onViewportHeightChange: ((CGFloat) -> Void)? = nil
+    var onPendingAssistantHeightChange: ((CGFloat) -> Void)? = nil
 
     private var hasPendingAssistant: Bool {
         messages.contains { message in
@@ -17,21 +20,19 @@ struct SybilChatTranscriptView: View {
     var body: some View {
         ScrollView {
             LazyVStack(alignment: .leading, spacing: 26) {
-                if isSending && !hasPendingAssistant {
-                    HStack(spacing: 8) {
-                        ProgressView()
-                            .controlSize(.small)
-                            .tint(SybilTheme.textMuted)
-                        Text("Assistant is typing…")
-                            .font(.sybil(.footnote))
-                            .foregroundStyle(SybilTheme.textMuted)
-                    }
-                    .scaleEffect(x: 1, y: -1)
-                }
-
                 ForEach(messages.reversed()) { message in
                     MessageBubble(message: message, isSending: isSending)
                         .frame(maxWidth: .infinity)
+                        .background {
+                            if isStreamingPendingAssistant(message) {
+                                GeometryReader { proxy in
+                                    Color.clear.preference(
+                                        key: SybilPendingAssistantHeightPreferenceKey.self,
+                                        value: proxy.size.height
+                                    )
+                                }
+                            }
+                        }
                         .scaleEffect(x: 1, y: -1)
                 }
 
@@ -45,13 +46,39 @@ struct SybilChatTranscriptView: View {
             }
             .frame(maxWidth: .infinity, alignment: .leading)
             .padding(.horizontal, 14)
-            .padding(.top, 18 + bottomContentInset)
+            .padding(.top, 18 + bottomContentInset + tailSpacerHeight)
             .padding(.bottom, 18 + topContentInset)
         }
         .frame(maxWidth: .infinity, alignment: .leading)
         .scrollDismissesKeyboard(.interactively)
+        .background {
+            GeometryReader { proxy in
+                Color.clear
+                    .onAppear {
+                        onViewportHeightChange?(proxy.size.height)
+                    }
+                    .onChange(of: proxy.size.height) { _, height in
+                        onViewportHeightChange?(height)
+                    }
+            }
+        }
+        .onPreferenceChange(SybilPendingAssistantHeightPreferenceKey.self) { height in
+            onPendingAssistantHeightChange?(height)
+        }
         .scaleEffect(x: 1, y: -1)
     }
+
+    private func isStreamingPendingAssistant(_ message: Message) -> Bool {
+        isSending && message.id.hasPrefix("temp-assistant-")
+    }
+}
+
+private struct SybilPendingAssistantHeightPreferenceKey: PreferenceKey {
+    static let defaultValue: CGFloat = 0
+
+    static func reduce(value: inout CGFloat, nextValue: () -> CGFloat) {
+        value = max(value, nextValue())
+    }
 }
 
 private struct MessageBubble: View {

ios/Packages/Sybil/Sources/Sybil/SybilModels.swift

@@ -4,12 +4,14 @@ public enum Provider: String, Codable, CaseIterable, Hashable, Sendable {
     case openai
     case anthropic
     case xai
+    case hermesAgent = "hermes-agent"
 
     public var displayName: String {
         switch self {
         case .openai: return "OpenAI"
         case .anthropic: return "Anthropic"
         case .xai: return "xAI"
+        case .hermesAgent: return "Hermes Agent"
         }
     }
 }
@@ -404,8 +406,8 @@ public struct CompletionRequestMessage: Codable, Sendable {
 }
 
 public struct CompletionStreamMeta: Codable, Sendable {
-    public var chatId: String
+    public var chatId: String?
-    public var callId: String
+    public var callId: String?
     public var provider: Provider
     public var model: String
 }

ios/Packages/Sybil/Sources/Sybil/SybilQuickActionRouting.swift

@@ -0,0 +1,19 @@
+import Combine
+import Foundation
+
+public enum SybilHomeScreenQuickAction {
+    public static let quickQuestionType = "net.buzzert.sybil2.quick-question"
+}
+
+@MainActor
+public final class SybilQuickActionRouter: ObservableObject {
+    public static let shared = SybilQuickActionRouter()
+
+    @Published public private(set) var quickQuestionPresentationRequest = 0
+
+    private init() {}
+
+    public func requestQuickQuestionPresentation() {
+        quickQuestionPresentationRequest += 1
+    }
+}

ios/Packages/Sybil/Sources/Sybil/SybilQuickQuestionView.swift

@@ -0,0 +1,297 @@
+import MarkdownUI
+import Observation
+import SwiftUI
+
+struct SybilQuickQuestionView: View {
+    @Bindable var viewModel: SybilViewModel
+    var focusRequest: Int
+
+    @Environment(\.dismiss) private var dismiss
+    @FocusState private var promptFocused: Bool
+
+    private var hasAnswerContent: Bool {
+        !viewModel.quickQuestionMessages.isEmpty || viewModel.quickQuestionError != nil
+    }
+
+    var body: some View {
+        VStack(spacing: 0) {
+            VStack(alignment: .leading, spacing: 16) {
+                header
+
+                answerArea
+
+                composer
+            }
+            .padding(.horizontal, 16)
+            .padding(.top, 18)
+            .padding(.bottom, 12)
+            .frame(maxWidth: 640, maxHeight: .infinity, alignment: .top)
+        }
+        .frame(maxWidth: .infinity, maxHeight: .infinity, alignment: .top)
+        .background(SybilTheme.backgroundGradient)
+        .preferredColorScheme(.dark)
+        .task(id: focusRequest) {
+            try? await Task.sleep(for: .milliseconds(260))
+            guard !Task.isCancelled else {
+                return
+            }
+            promptFocused = true
+        }
+    }
+
+    private var header: some View {
+        HStack {
+            Image(systemName: "sparkles")
+                .font(.system(size: 21, weight: .semibold))
+                .foregroundStyle(SybilTheme.primary)
+            
+            Text("Quick question")
+                .font(.title3.weight(.semibold))
+                .foregroundStyle(SybilTheme.text)
+                .lineLimit(1)
+            
+        }
+        .frame(maxWidth: .infinity, alignment: .leading)
+    }
+
+    private var answerArea: some View {
+        ScrollView {
+            VStack(alignment: .leading, spacing: 12) {
+                if hasAnswerContent {
+                    ForEach(viewModel.quickQuestionMessages) { message in
+                        QuickQuestionMessageView(message: message, isSending: viewModel.isQuickQuestionSending)
+                    }
+
+                    if let error = viewModel.quickQuestionError {
+                        Text(error)
+                            .font(.caption)
+                            .foregroundStyle(SybilTheme.danger)
+                            .fixedSize(horizontal: false, vertical: true)
+                    }
+                }
+            }
+            .frame(maxWidth: .infinity, alignment: .topLeading)
+            .padding(14)
+        }
+        .scrollDismissesKeyboard(.interactively)
+        .frame(maxWidth: .infinity, maxHeight: .infinity, alignment: .topLeading)
+        .background(
+            RoundedRectangle(cornerRadius: 12)
+                .fill(Color.black.opacity(0.36))
+        )
+        .overlay(
+            RoundedRectangle(cornerRadius: 12)
+                .stroke(SybilTheme.border.opacity(0.55), lineWidth: 1)
+        )
+    }
+
+    private var composer: some View {
+        VStack(alignment: .leading, spacing: 10) {
+            HStack(alignment: .bottom, spacing: 10) {
+                TextField(
+                    "Ask anything...",
+                    text: Binding(
+                        get: { viewModel.quickQuestionPrompt },
+                        set: { viewModel.updateQuickQuestionPrompt($0) }
+                    ),
+                    axis: .vertical
+                )
+                .focused($promptFocused)
+                .font(.body)
+                .textInputAutocapitalization(.sentences)
+                .autocorrectionDisabled(false)
+                .lineLimit(1 ... 6)
+                .submitLabel(.send)
+                .onSubmit(submitQuestion)
+                .padding(.horizontal, 12)
+                .padding(.vertical, 10)
+                .background(
+                    RoundedRectangle(cornerRadius: 12)
+                        .fill(SybilTheme.composerGradient)
+                        .opacity(0.98)
+                )
+                .foregroundStyle(SybilTheme.text)
+
+                Button(action: submitQuestion) {
+                    Image(systemName: "arrow.up")
+                        .font(.body.weight(.semibold))
+                        .frame(width: 40, height: 40)
+                        .background(
+                            Circle()
+                                .fill(
+                                    viewModel.canSendQuickQuestion
+                                        ? AnyShapeStyle(SybilTheme.primaryGradient)
+                                        : AnyShapeStyle(SybilTheme.surfaceStrong.opacity(0.92))
+                                )
+                        )
+                        .foregroundStyle(viewModel.canSendQuickQuestion ? SybilTheme.text : SybilTheme.textMuted)
+                }
+                .buttonStyle(.plain)
+                .disabled(!viewModel.canSendQuickQuestion)
+                .accessibilityLabel("Ask quick question")
+            }
+
+            controlsRow
+        }
+    }
+
+    private var convertButton: some View {
+        Button {
+            Task {
+                let didConvert = await viewModel.convertQuickQuestionToChat()
+                if didConvert {
+                    dismiss()
+                }
+            }
+        } label: {
+            Label("Chat", systemImage: "bubble.left")
+                .font(.caption.weight(.medium))
+                .lineLimit(1)
+                .minimumScaleFactor(0.8)
+        }
+        .buttonStyle(.plain)
+        .foregroundStyle(viewModel.canConvertQuickQuestion ? SybilTheme.text : SybilTheme.textMuted)
+        .padding(.horizontal, 10)
+        .frame(maxWidth: .infinity, minHeight: 40)
+        .background(
+            RoundedRectangle(cornerRadius: 12)
+                .fill(SybilTheme.surfaceStrong.opacity(0.78))
+                .overlay(
+                    RoundedRectangle(cornerRadius: 12)
+                        .stroke(SybilTheme.border.opacity(0.78), lineWidth: 1)
+                )
+        )
+        .disabled(!viewModel.canConvertQuickQuestion)
+    }
+
+    private var controlsRow: some View {
+        HStack(alignment: .center, spacing: 10) {
+            providerMenu
+            modelMenu
+            convertButton
+        }
+    }
+
+    private var providerMenu: some View {
+        Menu {
+            ForEach(viewModel.providerOptions, id: \.self) { provider in
+                Button {
+                    viewModel.setQuickQuestionProvider(provider)
+                } label: {
+                    if viewModel.quickQuestionProvider == provider {
+                        Label(provider.displayName, systemImage: "checkmark")
+                    } else {
+                        Text(provider.displayName)
+                    }
+                }
+            }
+        } label: {
+            QuickQuestionPickerPill(title: viewModel.quickQuestionProvider.displayName)
+        }
+        .frame(maxWidth: .infinity)
+        .disabled(viewModel.isQuickQuestionSending || viewModel.isConvertingQuickQuestion)
+        .accessibilityLabel("Quick question provider")
+    }
+
+    private var modelMenu: some View {
+        Menu {
+            if viewModel.quickQuestionProviderModelOptions.isEmpty {
+                Text("No models")
+            } else {
+                ForEach(viewModel.quickQuestionProviderModelOptions, id: \.self) { model in
+                    Button {
+                        viewModel.setQuickQuestionModel(model)
+                    } label: {
+                        if viewModel.quickQuestionModel == model {
+                            Label(model, systemImage: "checkmark")
+                        } else {
+                            Text(model)
+                        }
+                    }
+                }
+            }
+        } label: {
+            QuickQuestionPickerPill(title: viewModel.quickQuestionModel.isEmpty ? "No model" : viewModel.quickQuestionModel)
+        }
+        .frame(maxWidth: .infinity)
+        .disabled(viewModel.isQuickQuestionSending || viewModel.isConvertingQuickQuestion)
+        .accessibilityLabel("Quick question model")
+    }
+
+    private func submitQuestion() {
+        _ = viewModel.sendQuickQuestion()
+    }
+}
+
+private struct QuickQuestionPickerPill: View {
+    var title: String
+
+    var body: some View {
+        HStack(spacing: 8) {
+            Text(title)
+                .font(.caption.weight(.medium))
+                .foregroundStyle(SybilTheme.text)
+                .lineLimit(1)
+                .minimumScaleFactor(0.8)
+
+            Image(systemName: "chevron.down")
+                .font(.caption.weight(.semibold))
+                .foregroundStyle(SybilTheme.textMuted)
+        }
+        .padding(.horizontal, 10)
+        .frame(maxWidth: .infinity, minHeight: 40)
+        .background(
+            RoundedRectangle(cornerRadius: 12)
+                .fill(SybilTheme.surfaceStrong.opacity(0.78))
+                .overlay(
+                    RoundedRectangle(cornerRadius: 12)
+                        .stroke(SybilTheme.border.opacity(0.78), lineWidth: 1)
+                )
+        )
+    }
+}
+
+private struct QuickQuestionMessageView: View {
+    var message: Message
+    var isSending: Bool
+
+    private var isPendingAssistant: Bool {
+        message.id.hasPrefix("temp-assistant-quick-") &&
+            isSending &&
+            message.content.trimmingCharacters(in: .whitespacesAndNewlines).isEmpty
+    }
+
+    var body: some View {
+        if let metadata = message.toolCallMetadata {
+            Text(toolCallSummary(for: metadata, fallbackContent: message.content))
+                .font(.caption)
+                .foregroundStyle(SybilTheme.textMuted)
+                .fixedSize(horizontal: false, vertical: true)
+        } else if isPendingAssistant {
+            HStack(spacing: 8) {
+                ProgressView()
+                    .controlSize(.small)
+                    .tint(SybilTheme.primary)
+                Text("Thinking...")
+                    .font(.caption)
+                    .foregroundStyle(SybilTheme.textMuted)
+            }
+        } else if !message.content.trimmingCharacters(in: .whitespacesAndNewlines).isEmpty {
+            Markdown(message.content)
+                .font(.body)
+                .tint(SybilTheme.primary)
+                .foregroundStyle(SybilTheme.text.opacity(0.96))
+                .textSelection(.enabled)
+        }
+    }
+
+    private func toolCallSummary(for metadata: ToolCallMetadata, fallbackContent: String) -> String {
+        if let summary = metadata.summary?.trimmingCharacters(in: .whitespacesAndNewlines), !summary.isEmpty {
+            return summary
+        }
+        if !fallbackContent.trimmingCharacters(in: .whitespacesAndNewlines).isEmpty {
+            return fallbackContent
+        }
+        return "Ran \(metadata.toolName ?? "tool")."
+    }
+}

ios/Packages/Sybil/Sources/Sybil/SybilSettingsStore.swift

@@ -11,6 +11,12 @@ final class SybilSettingsStore {
         static let preferredOpenAIModel = "sybil.ios.preferredOpenAIModel"
         static let preferredAnthropicModel = "sybil.ios.preferredAnthropicModel"
         static let preferredXAIModel = "sybil.ios.preferredXAIModel"
+        static let preferredHermesAgentModel = "sybil.ios.preferredHermesAgentModel"
+        static let quickQuestionPreferredProvider = "sybil.ios.quickQuestionPreferredProvider"
+        static let quickQuestionPreferredOpenAIModel = "sybil.ios.quickQuestionPreferredOpenAIModel"
+        static let quickQuestionPreferredAnthropicModel = "sybil.ios.quickQuestionPreferredAnthropicModel"
+        static let quickQuestionPreferredXAIModel = "sybil.ios.quickQuestionPreferredXAIModel"
+        static let quickQuestionPreferredHermesAgentModel = "sybil.ios.quickQuestionPreferredHermesAgentModel"
     }
 
     private let defaults: UserDefaults
@@ -19,6 +25,8 @@ final class SybilSettingsStore {
     var adminToken: String
     var preferredProvider: Provider
     var preferredModelByProvider: [Provider: String]
+    var quickQuestionPreferredProvider: Provider
+    var quickQuestionPreferredModelByProvider: [Provider: String]
 
     init(defaults: UserDefaults = .standard) {
         self.defaults = defaults
@@ -32,10 +40,21 @@ final class SybilSettingsStore {
         let provider = defaults.string(forKey: Keys.preferredProvider).flatMap(Provider.init(rawValue:)) ?? .openai
         self.preferredProvider = provider
 
-        self.preferredModelByProvider = [
+        let preferredModels: [Provider: String] = [
             .openai: defaults.string(forKey: Keys.preferredOpenAIModel) ?? "gpt-4.1-mini",
             .anthropic: defaults.string(forKey: Keys.preferredAnthropicModel) ?? "claude-3-5-sonnet-latest",
-            .xai: defaults.string(forKey: Keys.preferredXAIModel) ?? "grok-3-mini"
+            .xai: defaults.string(forKey: Keys.preferredXAIModel) ?? "grok-3-mini",
+            .hermesAgent: defaults.string(forKey: Keys.preferredHermesAgentModel) ?? "hermes-agent"
+        ]
+        self.preferredModelByProvider = preferredModels
+
+        self.quickQuestionPreferredProvider =
+            defaults.string(forKey: Keys.quickQuestionPreferredProvider).flatMap(Provider.init(rawValue:)) ?? provider
+        self.quickQuestionPreferredModelByProvider = [
+            .openai: defaults.string(forKey: Keys.quickQuestionPreferredOpenAIModel) ?? preferredModels[.openai] ?? "gpt-4.1-mini",
+            .anthropic: defaults.string(forKey: Keys.quickQuestionPreferredAnthropicModel) ?? preferredModels[.anthropic] ?? "claude-3-5-sonnet-latest",
+            .xai: defaults.string(forKey: Keys.quickQuestionPreferredXAIModel) ?? preferredModels[.xai] ?? "grok-3-mini",
+            .hermesAgent: defaults.string(forKey: Keys.quickQuestionPreferredHermesAgentModel) ?? preferredModels[.hermesAgent] ?? "hermes-agent"
         ]
     }
 
@@ -53,6 +72,13 @@ final class SybilSettingsStore {
         defaults.set(preferredModelByProvider[.openai], forKey: Keys.preferredOpenAIModel)
         defaults.set(preferredModelByProvider[.anthropic], forKey: Keys.preferredAnthropicModel)
         defaults.set(preferredModelByProvider[.xai], forKey: Keys.preferredXAIModel)
+        defaults.set(preferredModelByProvider[.hermesAgent], forKey: Keys.preferredHermesAgentModel)
+
+        defaults.set(quickQuestionPreferredProvider.rawValue, forKey: Keys.quickQuestionPreferredProvider)
+        defaults.set(quickQuestionPreferredModelByProvider[.openai], forKey: Keys.quickQuestionPreferredOpenAIModel)
+        defaults.set(quickQuestionPreferredModelByProvider[.anthropic], forKey: Keys.quickQuestionPreferredAnthropicModel)
+        defaults.set(quickQuestionPreferredModelByProvider[.xai], forKey: Keys.quickQuestionPreferredXAIModel)
+        defaults.set(quickQuestionPreferredModelByProvider[.hermesAgent], forKey: Keys.quickQuestionPreferredHermesAgentModel)
     }
 
     var trimmedTokenOrNil: String? {
@@ -68,7 +94,7 @@ final class SybilSettingsStore {
             raw.removeLast()
         }
 
-        guard var components = URLComponents(string: raw) else {
+        guard let components = URLComponents(string: raw) else {
             return nil
         }
 

ios/Packages/Sybil/Sources/Sybil/SybilViewModel.swift

@@ -111,6 +111,16 @@ final class SybilViewModel {
     var provider: Provider
     var modelCatalog: [Provider: ProviderModelInfo] = [:]
     var model: String
+    var quickQuestionPrompt = ""
+    var quickQuestionMessages: [Message] = []
+    var quickQuestionError: String?
+    var quickQuestionProvider: Provider
+    var quickQuestionModel: String
+    var quickQuestionSubmittedPrompt: String?
+    var quickQuestionSubmittedProvider: Provider?
+    var quickQuestionSubmittedModel: String?
+    var isQuickQuestionSending = false
+    var isConvertingQuickQuestion = false
 
     @ObservationIgnored
     private var hasBootstrapped = false
@@ -132,6 +142,10 @@ final class SybilViewModel {
     @ObservationIgnored
     private var activeSearchAttachTasks: [String: Task<Void, Never>] = [:]
     @ObservationIgnored
+    private var quickQuestionTask: Task<Void, Never>?
+    @ObservationIgnored
+    private var quickQuestionRunID: UUID?
+    @ObservationIgnored
     private var isAppActive = true
     @ObservationIgnored
     private var appLifecycleGeneration = 0
@@ -141,7 +155,8 @@ final class SybilViewModel {
     private let fallbackModels: [Provider: [String]] = [
         .openai: ["gpt-4.1-mini"],
         .anthropic: ["claude-3-5-sonnet-latest"],
-        .xai: ["grok-3-mini"]
+        .xai: ["grok-3-mini"],
+        .hermesAgent: ["hermes-agent"]
     ]
 
     init(
@@ -152,14 +167,56 @@ final class SybilViewModel {
     ) {
         self.settings = settings
         self.clientFactory = clientFactory
-        self.provider = settings.preferredProvider
+        let initialProvider = settings.preferredProvider
-        self.model = settings.preferredModelByProvider[settings.preferredProvider] ?? "gpt-4.1-mini"
+        let initialModel = settings.preferredModelByProvider[initialProvider] ?? "gpt-4.1-mini"
+        self.provider = initialProvider
+        self.model = initialModel
+        let initialQuickQuestionProvider = settings.quickQuestionPreferredProvider
+        let initialQuickQuestionModel = settings.quickQuestionPreferredModelByProvider[initialQuickQuestionProvider] ?? initialModel
+        self.quickQuestionProvider = initialQuickQuestionProvider
+        self.quickQuestionModel = initialQuickQuestionModel
     }
 
     var providerModelOptions: [String] {
         modelOptions(for: provider)
     }
 
+    var providerOptions: [Provider] {
+        Provider.allCases.filter { candidate in
+            candidate != .hermesAgent || modelCatalog[candidate] != nil
+        }
+    }
+
+    var quickQuestionProviderModelOptions: [String] {
+        modelOptions(for: quickQuestionProvider)
+    }
+
+    var canSendQuickQuestion: Bool {
+        !isQuickQuestionSending &&
+            !isConvertingQuickQuestion &&
+            !quickQuestionPrompt.trimmingCharacters(in: .whitespacesAndNewlines).isEmpty &&
+            !quickQuestionModel.trimmingCharacters(in: .whitespacesAndNewlines).isEmpty
+    }
+
+    var quickQuestionAnswerText: String {
+        for message in quickQuestionMessages.reversed() where message.role == .assistant {
+            let content = message.content.trimmingCharacters(in: .whitespacesAndNewlines)
+            if !content.isEmpty {
+                return content
+            }
+        }
+        return ""
+    }
+
+    var canConvertQuickQuestion: Bool {
+        !isQuickQuestionSending &&
+            !isConvertingQuickQuestion &&
+            !(quickQuestionSubmittedPrompt?.trimmingCharacters(in: .whitespacesAndNewlines).isEmpty ?? true) &&
+            !quickQuestionAnswerText.isEmpty &&
+            quickQuestionSubmittedProvider != nil &&
+            !(quickQuestionSubmittedModel?.trimmingCharacters(in: .whitespacesAndNewlines).isEmpty ?? true)
+    }
+
     func modelOptions(for candidate: Provider) -> [String] {
         let serverModels = modelCatalog[candidate]?.models ?? []
         if !serverModels.isEmpty {
@@ -415,6 +472,7 @@ final class SybilViewModel {
         localActiveSearchIDs = []
         serverActiveChatIDs = []
         serverActiveSearchIDs = []
+        resetQuickQuestion()
         draftIdentity = UUID()
         composerAttachments = []
         settings.persist()
@@ -487,6 +545,159 @@ final class SybilViewModel {
         SybilLog.info(SybilLog.ui, "Provider changed to \(nextProvider.rawValue), model=\(nextModel)")
     }
 
+    func setQuickQuestionProvider(_ nextProvider: Provider) {
+        quickQuestionProvider = nextProvider
+
+        let options = modelOptions(for: nextProvider)
+        if let preferred = settings.quickQuestionPreferredModelByProvider[nextProvider], options.contains(preferred) {
+            quickQuestionModel = preferred
+        } else if let first = options.first {
+            quickQuestionModel = first
+        } else {
+            quickQuestionModel = ""
+        }
+
+        persistQuickQuestionModelSelection()
+    }
+
+    func setQuickQuestionModel(_ nextModel: String) {
+        quickQuestionModel = nextModel
+        persistQuickQuestionModelSelection()
+    }
+
+    private func persistQuickQuestionModelSelection() {
+        settings.quickQuestionPreferredProvider = quickQuestionProvider
+        let trimmedModel = quickQuestionModel.trimmingCharacters(in: .whitespacesAndNewlines)
+        if !trimmedModel.isEmpty {
+            settings.quickQuestionPreferredModelByProvider[quickQuestionProvider] = trimmedModel
+        }
+        settings.persist()
+    }
+
+    func updateQuickQuestionPrompt(_ nextPrompt: String) {
+        guard nextPrompt != quickQuestionPrompt else {
+            return
+        }
+
+        if isQuickQuestionSending || quickQuestionSubmittedPrompt != nil || !quickQuestionMessages.isEmpty {
+            cancelQuickQuestion()
+            quickQuestionSubmittedPrompt = nil
+            quickQuestionSubmittedProvider = nil
+            quickQuestionSubmittedModel = nil
+            quickQuestionMessages = []
+            quickQuestionError = nil
+        }
+
+        quickQuestionPrompt = nextPrompt
+    }
+
+    func resetQuickQuestion() {
+        cancelQuickQuestion()
+        quickQuestionPrompt = ""
+        quickQuestionMessages = []
+        quickQuestionError = nil
+        quickQuestionSubmittedPrompt = nil
+        quickQuestionSubmittedProvider = nil
+        quickQuestionSubmittedModel = nil
+        isConvertingQuickQuestion = false
+    }
+
+    func cancelQuickQuestion() {
+        quickQuestionTask?.cancel()
+        quickQuestionTask = nil
+        quickQuestionRunID = nil
+        isQuickQuestionSending = false
+    }
+
+    @discardableResult
+    func sendQuickQuestion() -> Task<Void, Never>? {
+        let content = quickQuestionPrompt.trimmingCharacters(in: .whitespacesAndNewlines)
+        guard !content.isEmpty, !isQuickQuestionSending, !isConvertingQuickQuestion else {
+            return nil
+        }
+
+        let selectedModel = quickQuestionModel.trimmingCharacters(in: .whitespacesAndNewlines)
+        guard !selectedModel.isEmpty else {
+            quickQuestionError = "No model available for selected provider."
+            return nil
+        }
+
+        cancelQuickQuestion()
+        let selectedProvider = quickQuestionProvider
+        let task = Task { [weak self] in
+            guard let self else {
+                return
+            }
+            await self.runQuickQuestion(prompt: content, provider: selectedProvider, model: selectedModel)
+        }
+        quickQuestionTask = task
+        return task
+    }
+
+    @discardableResult
+    func convertQuickQuestionToChat() async -> Bool {
+        let question = quickQuestionSubmittedPrompt?.trimmingCharacters(in: .whitespacesAndNewlines) ?? ""
+        let answer = quickQuestionAnswerText
+        guard !question.isEmpty,
+              !answer.isEmpty,
+              let submittedProvider = quickQuestionSubmittedProvider,
+              let submittedModel = quickQuestionSubmittedModel?.trimmingCharacters(in: .whitespacesAndNewlines),
+              !submittedModel.isEmpty,
+              !isQuickQuestionSending,
+              !isConvertingQuickQuestion
+        else {
+            return false
+        }
+
+        isConvertingQuickQuestion = true
+        quickQuestionError = nil
+        defer {
+            isConvertingQuickQuestion = false
+        }
+
+        do {
+            let titleSeed = question.split(whereSeparator: \.isNewline).first.map(String.init) ?? question
+            let title = String(titleSeed.trimmingCharacters(in: .whitespacesAndNewlines).prefix(48))
+            let chat = try await client().createChat(
+                title: title.isEmpty ? "Quick question" : title,
+                provider: submittedProvider,
+                model: submittedModel,
+                messages: [
+                    CompletionRequestMessage(role: .user, content: question),
+                    CompletionRequestMessage(role: .assistant, content: answer)
+                ]
+            )
+
+            setProvider(submittedProvider, model: submittedModel)
+            chats.removeAll(where: { $0.id == chat.id })
+            chats.insert(chat, at: 0)
+            draftKind = nil
+            selectedItem = .chat(chat.id)
+            selectedChat = ChatDetail(
+                id: chat.id,
+                title: chat.title,
+                createdAt: chat.createdAt,
+                updatedAt: chat.updatedAt,
+                initiatedProvider: chat.initiatedProvider,
+                initiatedModel: chat.initiatedModel,
+                lastUsedProvider: chat.lastUsedProvider,
+                lastUsedModel: chat.lastUsedModel,
+                messages: []
+            )
+            selectedSearch = nil
+            composer = ""
+            composerAttachments = []
+
+            await refreshCollections(preferredSelection: .chat(chat.id))
+            resetQuickQuestion()
+            return true
+        } catch {
+            quickQuestionError = normalizeAPIError(error)
+            SybilLog.error(SybilLog.ui, "Convert quick question to chat failed", error: error)
+            return false
+        }
+    }
+
     func startNewChat() {
         SybilLog.debug(SybilLog.ui, "Starting draft chat")
         resetSelectionLoading()
@@ -830,6 +1041,91 @@ final class SybilViewModel {
         isCreatingSearchChat = false
     }
 
+    private func runQuickQuestion(prompt: String, provider: Provider, model: String) async {
+        let runID = UUID()
+        quickQuestionRunID = runID
+        quickQuestionError = nil
+        quickQuestionSubmittedPrompt = prompt
+        quickQuestionSubmittedProvider = provider
+        quickQuestionSubmittedModel = model
+        quickQuestionMessages = [
+            Message(
+                id: "temp-assistant-quick-\(UUID().uuidString)",
+                createdAt: Date(),
+                role: .assistant,
+                content: "",
+                name: nil
+            )
+        ]
+        isQuickQuestionSending = true
+
+        defer {
+            if quickQuestionRunID == runID {
+                quickQuestionTask = nil
+                quickQuestionRunID = nil
+                isQuickQuestionSending = false
+            }
+        }
+
+        let streamStatus = CompletionStreamStatus()
+
+        do {
+            try await client().runCompletionStream(
+                body: CompletionStreamRequest(
+                    chatId: nil,
+                    persist: false,
+                    provider: provider,
+                    model: model,
+                    messages: [CompletionRequestMessage(role: .user, content: prompt)]
+                )
+            ) { [weak self] event in
+                guard let self else { return }
+                await self.applyQuickQuestionCompletionEvent(event, streamStatus: streamStatus)
+            }
+
+            if let streamError = await streamStatus.error() {
+                throw APIError.httpError(statusCode: 502, message: streamError)
+            }
+        } catch {
+            guard quickQuestionRunID == runID else {
+                return
+            }
+            if isCancellation(error) {
+                return
+            }
+
+            quickQuestionError = normalizeAPIError(error)
+            SybilLog.error(SybilLog.ui, "Quick question failed", error: error)
+        }
+    }
+
+    private func applyQuickQuestionCompletionEvent(_ event: CompletionStreamEvent, streamStatus: CompletionStreamStatus) async {
+        switch event {
+        case .meta:
+            break
+
+        case let .toolCall(payload):
+            insertQuickQuestionToolCallMessage(payload)
+
+        case let .delta(payload):
+            guard !payload.text.isEmpty else { return }
+            mutateQuickQuestionAssistantMessage { existing in
+                existing + payload.text
+            }
+
+        case let .done(payload):
+            mutateQuickQuestionAssistantMessage { _ in
+                payload.text
+            }
+
+        case let .error(payload):
+            await streamStatus.setError(payload.message)
+
+        case .ignored:
+            break
+        }
+    }
+
     private func loadInitialData(using client: any SybilAPIClienting) async {
         isLoadingCollections = true
         errorMessage = nil
@@ -893,6 +1189,11 @@ final class SybilViewModel {
     }
 
     private func syncModelSelectionWithServerCatalog() {
+        if !providerOptions.contains(provider), let firstProvider = providerOptions.first {
+            provider = firstProvider
+            settings.preferredProvider = firstProvider
+        }
+
         if !providerModelOptions.contains(model), let first = providerModelOptions.first {
             model = first
             settings.preferredModelByProvider[provider] = first
@@ -902,6 +1203,22 @@ final class SybilViewModel {
             model = preferred
         }
 
+        if !providerOptions.contains(quickQuestionProvider), let firstProvider = providerOptions.first {
+            quickQuestionProvider = firstProvider
+            settings.quickQuestionPreferredProvider = firstProvider
+        }
+
+        if !quickQuestionProviderModelOptions.contains(quickQuestionModel), let first = quickQuestionProviderModelOptions.first {
+            quickQuestionModel = first
+            settings.quickQuestionPreferredModelByProvider[quickQuestionProvider] = first
+        }
+
+        if let preferred = settings.quickQuestionPreferredModelByProvider[quickQuestionProvider],
+           quickQuestionProviderModelOptions.contains(preferred)
+        {
+            quickQuestionModel = preferred
+        }
+
         settings.persist()
     }
 
@@ -1752,6 +2069,15 @@ final class SybilViewModel {
         pendingChatStates[chatID] = pending
     }
 
+    private func mutateQuickQuestionAssistantMessage(_ transform: (String) -> String) {
+        let index = quickQuestionMessages.indices.last { quickQuestionMessages[$0].id.hasPrefix("temp-assistant-quick-") }
+        guard let index else {
+            return
+        }
+
+        quickQuestionMessages[index].content = transform(quickQuestionMessages[index].content)
+    }
+
     private func insertPendingToolCallMessage(_ payload: CompletionStreamToolCall, chatID: String) {
         guard var pending = pendingChatStates[chatID] else {
             return
@@ -1761,6 +2087,31 @@ final class SybilViewModel {
             return
         }
 
+        let message = toolCallMessage(for: payload)
+
+        if let assistantIndex = pending.messages.indices.last(where: { pending.messages[$0].id.hasPrefix("temp-assistant-") }) {
+            pending.messages.insert(message, at: assistantIndex)
+        } else {
+            pending.messages.append(message)
+        }
+
+        pendingChatStates[chatID] = pending
+    }
+
+    private func insertQuickQuestionToolCallMessage(_ payload: CompletionStreamToolCall) {
+        if quickQuestionMessages.contains(where: { $0.toolCallMetadata?.toolCallId == payload.toolCallId }) {
+            return
+        }
+
+        let message = toolCallMessage(for: payload)
+        if let assistantIndex = quickQuestionMessages.indices.last(where: { quickQuestionMessages[$0].id.hasPrefix("temp-assistant-quick-") }) {
+            quickQuestionMessages.insert(message, at: assistantIndex)
+        } else {
+            quickQuestionMessages.append(message)
+        }
+    }
+
+    private func toolCallMessage(for payload: CompletionStreamToolCall) -> Message {
         let metadata: JSONValue = .object([
             "kind": .string("tool_call"),
             "toolCallId": .string(payload.toolCallId),
@@ -1779,7 +2130,7 @@ final class SybilViewModel {
             ? "Ran tool '\(payload.name)'."
             : payload.summary
 
-        let message = Message(
+        return Message(
             id: "temp-tool-\(payload.toolCallId)",
             createdAt: Date(),
             role: .tool,
@@ -1787,14 +2138,6 @@ final class SybilViewModel {
             name: payload.name,
             metadata: metadata
         )
-
-        if let assistantIndex = pending.messages.indices.last(where: { pending.messages[$0].id.hasPrefix("temp-assistant-") }) {
-            pending.messages.insert(message, at: assistantIndex)
-        } else {
-            pending.messages.append(message)
-        }
-
-        pendingChatStates[chatID] = pending
     }
 
     private var currentChatID: String? {

ios/Packages/Sybil/Sources/Sybil/SybilWorkspaceView.swift

@@ -26,6 +26,10 @@ struct SybilWorkspaceView: View {
     @State private var isShowingPhotoPicker = false
     @State private var photoPickerItems: [PhotosPickerItem] = []
     @State private var isComposerDropTargeted = false
+    @State private var transcriptTailSpacerHeight = SybilTranscriptTailSpacer.minimumHeight
+    @State private var transcriptTailSpacerTargetHeight = SybilTranscriptTailSpacer.minimumHeight
+    @State private var transcriptViewportHeight: CGFloat = 0
+    @State private var pendingAssistantBaselineHeight: CGFloat?
     @State private var newChatSwipeOffset: CGFloat = 0
     @State private var newChatSwipeCompletionOffset: CGFloat = 0
     @State private var newChatSwipeContainerWidth: CGFloat = NewChatSwipeMetrics.referenceWidth
@@ -38,6 +42,10 @@ struct SybilWorkspaceView: View {
     private let customWorkspaceNavigationContentInset: CGFloat = 96
     private let composerOverlayContentInset: CGFloat = 112
 
+    private var visibleTranscriptTailSpacerHeight: CGFloat {
+        viewModel.showsComposer && !viewModel.isSearchMode ? transcriptTailSpacerHeight : 0
+    }
+
     private var isSettingsSelected: Bool {
         if case .settings = viewModel.selectedItem {
             return true
@@ -145,6 +153,17 @@ struct SybilWorkspaceView: View {
             }
             resetNewChatSwipe(animated: false)
         }
+        .onChange(of: transcriptScrollContextID) { _, _ in
+            handleTranscriptContextChange()
+        }
+        .onChange(of: viewModel.isSendingVisibleChat) { wasSending, isSending in
+            handleVisibleChatSendingChange(wasSending: wasSending, isSending: isSending)
+        }
+        .onChange(of: viewModel.errorMessage) { _, message in
+            if message != nil && !viewModel.isSendingVisibleChat {
+                resetTranscriptTailSpacer(animated: true)
+            }
+        }
         .task(id: composerFocusPolicyID) {
             await applyComposerFocusPolicy()
         }
@@ -194,7 +213,14 @@ struct SybilWorkspaceView: View {
                         isLoading: viewModel.isLoadingSelection,
                         isSending: viewModel.isSendingVisibleChat,
                         topContentInset: showsCustomWorkspaceNavigation ? customWorkspaceNavigationContentInset : 0,
-                        bottomContentInset: viewModel.showsComposer ? composerOverlayContentInset : 0
+                        bottomContentInset: viewModel.showsComposer ? composerOverlayContentInset : 0,
+                        tailSpacerHeight: visibleTranscriptTailSpacerHeight,
+                        onViewportHeightChange: { height in
+                            handleTranscriptViewportHeightChange(height)
+                        },
+                        onPendingAssistantHeightChange: { height in
+                            handlePendingAssistantHeightChange(height)
+                        }
                     )
                     .id(transcriptScrollContextID)
                 }
@@ -285,6 +311,86 @@ struct SybilWorkspaceView: View {
         }
     }
 
+    private func handleTranscriptContextChange() {
+        resetTranscriptTailSpacer(animated: false)
+    }
+
+    private func handleVisibleChatSendingChange(wasSending: Bool, isSending: Bool) {
+        guard !viewModel.isSearchMode else {
+            resetTranscriptTailSpacer(animated: true)
+            return
+        }
+
+        if isSending {
+            prepareTranscriptTailSpacerForReply(animated: false)
+            return
+        }
+
+        if wasSending {
+            if viewModel.errorMessage != nil {
+                resetTranscriptTailSpacer(animated: true)
+            }
+            pendingAssistantBaselineHeight = nil
+        }
+    }
+
+    private func handleTranscriptViewportHeightChange(_ height: CGFloat) {
+        transcriptViewportHeight = height
+
+        if viewModel.isSendingVisibleChat,
+           transcriptTailSpacerTargetHeight <= SybilTranscriptTailSpacer.minimumHeight {
+            prepareTranscriptTailSpacerForReply(animated: false)
+        }
+    }
+
+    private func handlePendingAssistantHeightChange(_ height: CGFloat) {
+        guard viewModel.isSendingVisibleChat, !viewModel.isSearchMode, height > 0 else {
+            return
+        }
+
+        if pendingAssistantBaselineHeight == nil {
+            pendingAssistantBaselineHeight = height
+        }
+
+        let measuredHeight = SybilTranscriptTailSpacer.placeholderHeight(
+            targetHeight: transcriptTailSpacerTargetHeight,
+            baselineAssistantHeight: pendingAssistantBaselineHeight ?? height,
+            currentAssistantHeight: height
+        )
+        let nextHeight = min(transcriptTailSpacerHeight, measuredHeight)
+        setTranscriptTailSpacer(nextHeight, animated: false)
+    }
+
+    private func prepareTranscriptTailSpacerForReply(animated: Bool) {
+        let targetHeight = SybilTranscriptTailSpacer.replyBufferHeight(for: transcriptViewportHeight)
+        transcriptTailSpacerTargetHeight = targetHeight
+        pendingAssistantBaselineHeight = nil
+        setTranscriptTailSpacer(targetHeight, animated: animated)
+    }
+
+    private func resetTranscriptTailSpacer(animated: Bool) {
+        transcriptTailSpacerTargetHeight = SybilTranscriptTailSpacer.minimumHeight
+        pendingAssistantBaselineHeight = nil
+        setTranscriptTailSpacer(SybilTranscriptTailSpacer.minimumHeight, animated: animated)
+    }
+
+    private func setTranscriptTailSpacer(_ height: CGFloat, animated: Bool) {
+        let nextHeight = SybilTranscriptTailSpacer.clampedHeight(height)
+        guard abs(nextHeight - transcriptTailSpacerHeight) >= 0.5 else {
+            return
+        }
+
+        let update = {
+            transcriptTailSpacerHeight = nextHeight
+        }
+
+        if animated {
+            withAnimation(.easeOut(duration: 0.22), update)
+        } else {
+            update()
+        }
+    }
+
     private func beginNewChatSwipe(containerWidth: CGFloat) {
         let update = {
             newChatSwipeContainerWidth = max(containerWidth, 1)
@@ -495,7 +601,7 @@ struct SybilWorkspaceView: View {
 
         Divider()
 
-        ForEach(Provider.allCases, id: \.self) { candidate in
+        ForEach(viewModel.providerOptions, id: \.self) { candidate in
             Menu(candidate.displayName) {
                 let models = viewModel.modelOptions(for: candidate)
                 if models.isEmpty {
@@ -702,6 +808,10 @@ struct SybilWorkspaceView: View {
             return
         }
 
+        if !viewModel.isSearchMode {
+            prepareTranscriptTailSpacerForReply(animated: false)
+        }
+
         #if !targetEnvironment(macCatalyst)
         if !viewModel.isSearchMode {
             composerFocused = false
@@ -771,6 +881,37 @@ struct SybilWorkspaceView: View {
     }
 }
 
+enum SybilTranscriptTailSpacer {
+    static let minimumHeight: CGFloat = 20
+    static let replyBufferMin: CGFloat = 288
+    static let replyBufferMax: CGFloat = 576
+    static let replyBufferViewportRatio: CGFloat = 0.52
+
+    static func replyBufferHeight(for viewportHeight: CGFloat) -> CGFloat {
+        guard viewportHeight > 0 else {
+            return replyBufferMin
+        }
+
+        return min(
+            replyBufferMax,
+            max(replyBufferMin, (viewportHeight * replyBufferViewportRatio).rounded())
+        )
+    }
+
+    static func clampedHeight(_ height: CGFloat) -> CGFloat {
+        max(minimumHeight, height.rounded(.up))
+    }
+
+    static func placeholderHeight(
+        targetHeight: CGFloat,
+        baselineAssistantHeight: CGFloat,
+        currentAssistantHeight: CGFloat
+    ) -> CGFloat {
+        let consumedHeight = max(currentAssistantHeight - baselineAssistantHeight, 0)
+        return clampedHeight(targetHeight - consumedHeight)
+    }
+}
+
 enum NewChatSwipeMetrics {
     static let referenceWidth: CGFloat = 390
     static let horizontalActivationDistance: CGFloat = 18

ios/Packages/Sybil/Tests/SybilTests/SybilTests.swift

@@ -6,13 +6,22 @@ import Testing
 private struct MockClientCallSnapshot: Sendable {
     var listChats = 0
     var listSearches = 0
+    var createChat = 0
     var getChat = 0
     var getSearch = 0
     var getActiveRuns = 0
+    var runCompletionStream = 0
     var attachCompletionStream = 0
     var attachSearchStream = 0
 }
 
+private struct ChatCreateCallSnapshot: Sendable {
+    var title: String?
+    var provider: Provider?
+    var model: String?
+    var messages: [CompletionRequestMessage]?
+}
+
 private struct UnexpectedClientCall: Error {}
 
 private actor MockSybilClient: SybilAPIClienting {
@@ -24,6 +33,9 @@ private actor MockSybilClient: SybilAPIClienting {
     private let activeRunsResponse: ActiveRunsResponse
 
     private var snapshot = MockClientCallSnapshot()
+    private var lastCreateChatCall: ChatCreateCallSnapshot?
+    private var lastCompletionStreamBody: CompletionStreamRequest?
+    private var completionStreamEvents: [CompletionStreamEvent]?
     private var getChatDelayNanoseconds: UInt64 = 0
     private var getSearchDelayNanoseconds: UInt64 = 0
     private var completionStreamNetworkErrorMessage: String?
@@ -55,6 +67,19 @@ private actor MockSybilClient: SybilAPIClienting {
         snapshot
     }
 
+    func currentCreateChatCall() -> ChatCreateCallSnapshot? {
+        lastCreateChatCall
+    }
+
+    func currentCompletionStreamBody() -> CompletionStreamRequest? {
+        lastCompletionStreamBody
+    }
+
+    func setCompletionStreamEvents(_ events: [CompletionStreamEvent], delayNanoseconds: UInt64 = 0) {
+        completionStreamEvents = events
+        completionStreamDelayNanoseconds = delayNanoseconds
+    }
+
     func setCompletionStreamNetworkError(_ message: String, delayNanoseconds: UInt64 = 0) {
         completionStreamNetworkErrorMessage = message
         completionStreamDelayNanoseconds = delayNanoseconds
@@ -100,7 +125,19 @@ private actor MockSybilClient: SybilAPIClienting {
         return chatsResponse
     }
 
-    func createChat(title: String?) async throws -> ChatSummary {
+    func createChat(
+        title: String?,
+        provider: Provider?,
+        model: String?,
+        messages: [CompletionRequestMessage]?
+    ) async throws -> ChatSummary {
+        snapshot.createChat += 1
+        lastCreateChatCall = ChatCreateCallSnapshot(
+            title: title,
+            provider: provider,
+            model: model,
+            messages: messages
+        )
         if let createChatResponse {
             return createChatResponse
         }
@@ -167,12 +204,20 @@ private actor MockSybilClient: SybilAPIClienting {
         body: CompletionStreamRequest,
         onEvent: @escaping @Sendable (CompletionStreamEvent) async -> Void
     ) async throws {
+        snapshot.runCompletionStream += 1
+        lastCompletionStreamBody = body
         if completionStreamDelayNanoseconds > 0 {
             try await Task.sleep(nanoseconds: completionStreamDelayNanoseconds)
         }
         if let completionStreamNetworkErrorMessage {
             throw APIError.networkError(message: completionStreamNetworkErrorMessage)
         }
+        if let completionStreamEvents {
+            for event in completionStreamEvents {
+                await onEvent(event)
+            }
+            return
+        }
         throw UnexpectedClientCall()
     }
 
@@ -470,6 +515,117 @@ private func makeSearchDetail(id: String, date: Date, answer: String) -> SearchD
     await sendTask.value
 }
 
+@MainActor
+@Test func quickQuestionRunsNonPersistentCompletionStream() async throws {
+    let client = MockSybilClient()
+    await client.setCompletionStreamEvents([
+        .delta(CompletionStreamDelta(text: "Reset it from ")),
+        .done(CompletionStreamDone(text: "Reset it from Settings."))
+    ])
+    let viewModel = SybilViewModel(settings: testSettings(named: #function)) { _ in client }
+    viewModel.isAuthenticated = true
+    viewModel.isCheckingSession = false
+    viewModel.quickQuestionPrompt = "How do I reset my password?"
+
+    let task = viewModel.sendQuickQuestion()
+    await task?.value
+
+    let snapshot = await client.currentSnapshot()
+    let body = await client.currentCompletionStreamBody()
+    #expect(snapshot.runCompletionStream == 1)
+    #expect(body?.persist == false)
+    #expect(body?.chatId == nil)
+    #expect(body?.provider == .openai)
+    #expect(body?.messages.first?.role == .user)
+    #expect(body?.messages.first?.content == "How do I reset my password?")
+    #expect(viewModel.quickQuestionAnswerText == "Reset it from Settings.")
+    #expect(!viewModel.isQuickQuestionSending)
+}
+
+@MainActor
+@Test func quickQuestionConvertCreatesSeededChat() async throws {
+    let date = Date(timeIntervalSince1970: 1_700_000_250)
+    let chat = makeChatSummary(id: "quick-chat", date: date)
+    let detail = ChatDetail(
+        id: chat.id,
+        title: chat.title,
+        createdAt: chat.createdAt,
+        updatedAt: chat.updatedAt,
+        initiatedProvider: .openai,
+        initiatedModel: "gpt-4.1-mini",
+        lastUsedProvider: .openai,
+        lastUsedModel: "gpt-4.1-mini",
+        messages: [
+            Message(id: "quick-user", createdAt: date, role: .user, content: "How do I reset my password?", name: nil),
+            Message(id: "quick-assistant", createdAt: date, role: .assistant, content: "Reset it from Settings.", name: nil)
+        ]
+    )
+    let client = MockSybilClient(
+        chatsResponse: [chat],
+        chatDetails: [chat.id: detail],
+        createChatResponse: chat
+    )
+    let viewModel = SybilViewModel(settings: testSettings(named: #function)) { _ in client }
+    viewModel.isAuthenticated = true
+    viewModel.isCheckingSession = false
+    viewModel.quickQuestionSubmittedPrompt = "How do I reset my password?"
+    viewModel.quickQuestionSubmittedProvider = .openai
+    viewModel.quickQuestionSubmittedModel = "gpt-4.1-mini"
+    viewModel.quickQuestionMessages = [
+        Message(
+            id: "temp-assistant-quick",
+            createdAt: date,
+            role: .assistant,
+            content: "Reset it from Settings.",
+            name: nil
+        )
+    ]
+
+    let didConvert = await viewModel.convertQuickQuestionToChat()
+
+    let snapshot = await client.currentSnapshot()
+    let createCall = await client.currentCreateChatCall()
+    #expect(didConvert)
+    #expect(snapshot.createChat == 1)
+    #expect(createCall?.title == "How do I reset my password?")
+    #expect(createCall?.provider == .openai)
+    #expect(createCall?.model == "gpt-4.1-mini")
+    #expect(createCall?.messages?.map(\.role) == [.user, .assistant])
+    #expect(createCall?.messages?.map(\.content) == ["How do I reset my password?", "Reset it from Settings."])
+    #expect(viewModel.selectedItem == .chat("quick-chat"))
+    #expect(viewModel.quickQuestionPrompt.isEmpty)
+}
+
+@MainActor
+@Test func quickQuestionProviderAndModelSelectionPersistSeparately() async throws {
+    let defaults = UserDefaults(suiteName: #function)!
+    defaults.removePersistentDomain(forName: #function)
+    let settings = SybilSettingsStore(defaults: defaults)
+    settings.apiBaseURL = "http://127.0.0.1:8787"
+    let viewModel = SybilViewModel(settings: settings) { _ in MockSybilClient() }
+    viewModel.modelCatalog = [
+        .openai: ProviderModelInfo(models: ["gpt-4.1-mini", "gpt-4o"], loadedAt: nil, error: nil),
+        .anthropic: ProviderModelInfo(models: ["claude-3-5-sonnet-latest", "claude-3-haiku"], loadedAt: nil, error: nil)
+    ]
+
+    viewModel.setQuickQuestionProvider(.anthropic)
+    viewModel.setQuickQuestionModel("claude-3-haiku")
+
+    #expect(viewModel.quickQuestionProvider == .anthropic)
+    #expect(viewModel.quickQuestionModel == "claude-3-haiku")
+    #expect(settings.preferredProvider == .openai)
+
+    let reloadedSettings = SybilSettingsStore(defaults: defaults)
+    #expect(reloadedSettings.quickQuestionPreferredProvider == .anthropic)
+    #expect(reloadedSettings.quickQuestionPreferredModelByProvider[.anthropic] == "claude-3-haiku")
+    #expect(reloadedSettings.preferredProvider == .openai)
+
+    let reloadedViewModel = SybilViewModel(settings: reloadedSettings) { _ in MockSybilClient() }
+    #expect(reloadedViewModel.quickQuestionProvider == .anthropic)
+    #expect(reloadedViewModel.quickQuestionModel == "claude-3-haiku")
+    #expect(reloadedViewModel.provider == .openai)
+}
+
 @MainActor
 @Test func reconnectAttachesSelectedActiveChatStream() async throws {
     let date = Date(timeIntervalSince1970: 1_700_000_260)
@@ -628,3 +784,23 @@ private func makeSearchDetail(id: String, date: Date, answer: String) -> SearchD
     #expect(BackSwipeMetrics.shouldComplete(offset: 24, velocityX: 800, width: width, isLatched: false))
     #expect(!BackSwipeMetrics.shouldComplete(offset: latchDistance + 1, velocityX: -800, width: width, isLatched: true))
 }
+
+@Test func transcriptTailSpacerContractsAsContentGrows() async throws {
+    let targetHeight: CGFloat = 320
+    let baselineAssistantHeight: CGFloat = 28
+
+    #expect(
+        SybilTranscriptTailSpacer.placeholderHeight(
+            targetHeight: targetHeight,
+            baselineAssistantHeight: baselineAssistantHeight,
+            currentAssistantHeight: baselineAssistantHeight + 120
+        ) == 200
+    )
+    #expect(
+        SybilTranscriptTailSpacer.placeholderHeight(
+            targetHeight: targetHeight,
+            baselineAssistantHeight: baselineAssistantHeight,
+            currentAssistantHeight: baselineAssistantHeight + 500
+        ) == SybilTranscriptTailSpacer.minimumHeight
+    )
+}

server/README.md

@@ -1,7 +1,7 @@
 # Sybil Server
 
 Backend API for:
-- LLM multiplexer (OpenAI Responses / Anthropic / xAI Chat Completions-compatible Grok)
+- LLM multiplexer (OpenAI Responses / Anthropic / xAI Chat Completions-compatible Grok / Hermes Agent)
 - Personal chat database (chats/messages + LLM call log)
 
 ## Stack
@@ -43,6 +43,9 @@ If `ADMIN_TOKEN` is not set, the server runs in open mode (dev).
 - `OPENAI_API_KEY`
 - `ANTHROPIC_API_KEY`
 - `XAI_API_KEY`
+- `HERMES_AGENT_API_BASE_URL` (`http://127.0.0.1:8642/v1` by default; include the `/v1` suffix)
+- `HERMES_AGENT_API_KEY` (enables the Hermes Agent provider; set to Hermes `API_SERVER_KEY`, or any non-empty value if that local server does not require auth)
+- `HERMES_AGENT_MODEL` (optional fallback/override model id; defaults client-side to `hermes-agent`)
 - `EXA_API_KEY`
 - `CHAT_WEB_SEARCH_ENGINE` (`exa` by default, or `searxng` for chat tool calls only)
 - `SEARXNG_BASE_URL` (required when `CHAT_WEB_SEARCH_ENGINE=searxng`; instance must allow `format=json`)

server/prisma/schema.prisma

@@ -13,6 +13,7 @@ enum Provider {
   openai
   anthropic
   xai
+  hermes_agent @map("hermes-agent")
 }
 
 enum MessageRole {

server/src/env.ts

@@ -11,6 +11,13 @@ const OptionalUrlSchema = z.preprocess(
   z.string().trim().url().optional()
 );
 
+const DEFAULT_HERMES_AGENT_API_BASE_URL = "http://127.0.0.1:8642/v1";
+
+const HermesAgentApiBaseUrlSchema = z.preprocess(
+  (value) => (typeof value === "string" && value.trim() === "" ? undefined : value),
+  z.string().trim().url().default(DEFAULT_HERMES_AGENT_API_BASE_URL)
+);
+
 const ChatWebSearchEngineSchema = z.preprocess(
   (value) => {
     if (typeof value !== "string") return value;
@@ -59,6 +66,9 @@ const EnvSchema = z.object({
   OPENAI_API_KEY: z.string().optional(),
   ANTHROPIC_API_KEY: z.string().optional(),
   XAI_API_KEY: z.string().optional(),
+  HERMES_AGENT_API_BASE_URL: HermesAgentApiBaseUrlSchema,
+  HERMES_AGENT_API_KEY: OptionalTrimmedStringSchema,
+  HERMES_AGENT_MODEL: OptionalTrimmedStringSchema,
   EXA_API_KEY: z.string().optional(),
 
   // Chat-mode web_search tool configuration. Search mode remains Exa-only for now.

server/src/llm/chat-tools.ts

@@ -385,6 +385,10 @@ function normalizeIncomingMessages(messages: ChatMessage[]) {
   return [{ role: "system", content: CHAT_TOOL_SYSTEM_PROMPT }, ...normalized];
 }
 
+function normalizePlainIncomingMessages(messages: ChatMessage[]) {
+  return messages.map((message) => buildOpenAIConversationMessage(message));
+}
+
 function normalizeIncomingResponsesInput(messages: ChatMessage[]) {
   const normalized = messages.map((message) => buildOpenAIResponsesInputMessage(message));
 
@@ -853,6 +857,20 @@ function extractResponsesText(response: any, fallback = "") {
   return parts.join("") || fallback;
 }
 
+function extractChatCompletionContent(message: any) {
+  if (typeof message?.content === "string") return message.content;
+  if (!Array.isArray(message?.content)) return "";
+
+  return message.content
+    .map((part: any) => {
+      if (typeof part === "string") return part;
+      if (typeof part?.text === "string") return part.text;
+      if (typeof part?.content === "string") return part.content;
+      return "";
+    })
+    .join("");
+}
+
 function getUnstreamedText(finalText: string, streamedText: string) {
   if (!finalText) return "";
   if (!streamedText) return finalText;
@@ -1093,6 +1111,26 @@ export async function runToolAwareChatCompletions(params: ToolAwareCompletionPar
   };
 }
 
+export async function runPlainChatCompletions(params: ToolAwareCompletionParams): Promise<ToolAwareCompletionResult> {
+  const completion = await params.client.chat.completions.create({
+    model: params.model,
+    messages: normalizePlainIncomingMessages(params.messages),
+    temperature: params.temperature,
+    max_tokens: params.maxTokens,
+  } as any);
+
+  const usageAcc: Required<ToolAwareUsage> = { inputTokens: 0, outputTokens: 0, totalTokens: 0 };
+  const sawUsage = mergeUsage(usageAcc, completion?.usage);
+  const message = completion?.choices?.[0]?.message;
+
+  return {
+    text: extractChatCompletionContent(message),
+    usage: sawUsage ? usageAcc : undefined,
+    raw: { response: completion, api: "chat.completions" },
+    toolEvents: [],
+  };
+}
+
 export async function* runToolAwareOpenAIChatStream(
   params: ToolAwareCompletionParams
 ): AsyncGenerator<ToolAwareStreamingEvent> {
@@ -1354,3 +1392,41 @@ export async function* runToolAwareChatCompletionsStream(
     },
   };
 }
+
+export async function* runPlainChatCompletionsStream(
+  params: ToolAwareCompletionParams
+): AsyncGenerator<ToolAwareStreamingEvent> {
+  const rawResponses: unknown[] = [];
+  const usageAcc: Required<ToolAwareUsage> = { inputTokens: 0, outputTokens: 0, totalTokens: 0 };
+  let sawUsage = false;
+  let text = "";
+
+  const stream = await params.client.chat.completions.create({
+    model: params.model,
+    messages: normalizePlainIncomingMessages(params.messages),
+    temperature: params.temperature,
+    max_tokens: params.maxTokens,
+    stream: true,
+  } as any);
+
+  for await (const chunk of stream as any as AsyncIterable<any>) {
+    rawResponses.push(chunk);
+    sawUsage = mergeUsage(usageAcc, chunk?.usage) || sawUsage;
+
+    const deltaText = chunk?.choices?.[0]?.delta?.content ?? "";
+    if (typeof deltaText === "string" && deltaText.length) {
+      text += deltaText;
+      yield { type: "delta", text: deltaText };
+    }
+  }
+
+  yield {
+    type: "done",
+    result: {
+      text,
+      usage: sawUsage ? usageAcc : undefined,
+      raw: { streamed: true, responses: rawResponses, api: "chat.completions" },
+      toolEvents: [],
+    },
+  };
+}

server/src/llm/model-catalog.ts

@@ -1,5 +1,6 @@
 import type { FastifyBaseLogger } from "fastify";
-import { anthropicClient, openaiClient, xaiClient } from "./providers.js";
+import { env } from "../env.js";
+import { anthropicClient, hermesAgentClient, isHermesAgentConfigured, openaiClient, xaiClient } from "./providers.js";
 import type { Provider } from "./types.js";
 
 export type ProviderModelSnapshot = {
@@ -8,9 +9,9 @@ export type ProviderModelSnapshot = {
   error: string | null;
 };
 
-export type ModelCatalogSnapshot = Record<Provider, ProviderModelSnapshot>;
+export type ModelCatalogSnapshot = Partial<Record<Provider, ProviderModelSnapshot>>;
 
-const providers: Provider[] = ["openai", "anthropic", "xai"];
+const baseProviders: Provider[] = ["openai", "anthropic", "xai"];
 const MODEL_FETCH_TIMEOUT_MS = 15000;
 
 const modelCatalog: ModelCatalogSnapshot = {
@@ -19,6 +20,10 @@ const modelCatalog: ModelCatalogSnapshot = {
   xai: { models: [], loadedAt: null, error: null },
 };
 
+function getCatalogProviders(): Provider[] {
+  return isHermesAgentConfigured() ? [...baseProviders, "hermes-agent"] : baseProviders;
+}
+
 function uniqSorted(models: string[]) {
   return [...new Set(models.map((value) => value.trim()).filter(Boolean))].sort((a, b) => a.localeCompare(b));
 }
@@ -59,10 +64,17 @@ async function fetchProviderModels(provider: Provider) {
     return uniqSorted(page.data.map((model) => model.id));
   }
 
+  if (provider === "xai") {
     const page = await xaiClient().models.list();
     return uniqSorted(page.data.map((model) => model.id));
   }
 
+  const page = await hermesAgentClient().models.list();
+  const models = page.data.map((model) => model.id);
+  if (env.HERMES_AGENT_MODEL) models.push(env.HERMES_AGENT_MODEL);
+  return uniqSorted(models);
+}
+
 async function refreshProviderModels(provider: Provider, logger?: FastifyBaseLogger) {
   try {
     const models = await withTimeout(fetchProviderModels(provider), MODEL_FETCH_TIMEOUT_MS, `${provider} model fetch`);
@@ -75,7 +87,7 @@ async function refreshProviderModels(provider: Provider, logger?: FastifyBaseLog
   } catch (err: any) {
     const message = err?.message ?? String(err);
     modelCatalog[provider] = {
-      models: [],
+      models: provider === "hermes-agent" && env.HERMES_AGENT_MODEL ? [env.HERMES_AGENT_MODEL] : [],
       loadedAt: new Date().toISOString(),
       error: message,
     };
@@ -84,25 +96,18 @@ async function refreshProviderModels(provider: Provider, logger?: FastifyBaseLog
 }
 
 export async function warmModelCatalog(logger?: FastifyBaseLogger) {
-  await Promise.all(providers.map((provider) => refreshProviderModels(provider, logger)));
+  await Promise.all(getCatalogProviders().map((provider) => refreshProviderModels(provider, logger)));
 }
 
 export function getModelCatalogSnapshot(): ModelCatalogSnapshot {
-  return {
+  const snapshot: ModelCatalogSnapshot = {};
-    openai: {
+  for (const provider of getCatalogProviders()) {
-      models: [...modelCatalog.openai.models],
+    const entry = modelCatalog[provider] ?? { models: [], loadedAt: null, error: null };
-      loadedAt: modelCatalog.openai.loadedAt,
+    snapshot[provider] = {
-      error: modelCatalog.openai.error,
+      models: [...entry.models],
-    },
+      loadedAt: entry.loadedAt,
-    anthropic: {
+      error: entry.error,
-      models: [...modelCatalog.anthropic.models],
-      loadedAt: modelCatalog.anthropic.loadedAt,
-      error: modelCatalog.anthropic.error,
-    },
-    xai: {
-      models: [...modelCatalog.xai.models],
-      loadedAt: modelCatalog.xai.loadedAt,
-      error: modelCatalog.xai.error,
-    },
     };
   }
+  return snapshot;
+}

server/src/llm/multiplexer.ts

@@ -1,13 +1,13 @@
 import { performance } from "node:perf_hooks";
 import { prisma } from "../db.js";
-import { anthropicClient, openaiClient, xaiClient } from "./providers.js";
+import { anthropicClient, hermesAgentClient, openaiClient, xaiClient } from "./providers.js";
-import { buildToolLogMessageData, runToolAwareChatCompletions, runToolAwareOpenAIChat } from "./chat-tools.js";
+import { buildToolLogMessageData, runPlainChatCompletions, runToolAwareChatCompletions, runToolAwareOpenAIChat } from "./chat-tools.js";
 import { buildAnthropicConversationMessage, getAnthropicSystemPrompt } from "./message-content.js";
+import { toPrismaProvider } from "./provider-ids.js";
 import type { MultiplexRequest, MultiplexResponse, Provider } from "./types.js";
 
 function asProviderEnum(p: Provider) {
-  // Prisma enum values match these strings.
+  return toPrismaProvider(p);
-  return p;
 }
 
 export async function runMultiplex(req: MultiplexRequest): Promise<MultiplexResponse> {
@@ -84,6 +84,23 @@ export async function runMultiplex(req: MultiplexRequest): Promise<MultiplexResp
       outText = r.text;
       usage = r.usage;
       toolMessages = r.toolEvents.map((event) => buildToolLogMessageData(call.chatId, event));
+    } else if (req.provider === "hermes-agent") {
+      const client = hermesAgentClient();
+      const r = await runPlainChatCompletions({
+        client,
+        model: req.model,
+        messages: req.messages,
+        temperature: req.temperature,
+        maxTokens: req.maxTokens,
+        logContext: {
+          provider: req.provider,
+          model: req.model,
+          chatId,
+        },
+      });
+      raw = r.raw;
+      outText = r.text;
+      usage = r.usage;
     } else if (req.provider === "anthropic") {
       const client = anthropicClient();
 

server/src/llm/provider-ids.ts

@@ -0,0 +1,31 @@
+import type { Provider } from "./types.js";
+
+type PrismaProvider = Exclude<Provider, "hermes-agent"> | "hermes_agent";
+
+export function toPrismaProvider(provider: Provider): PrismaProvider {
+  return provider === "hermes-agent" ? "hermes_agent" : provider;
+}
+
+export function fromPrismaProvider(provider: unknown): Provider | null {
+  if (provider === null || provider === undefined) return null;
+  if (provider === "hermes_agent" || provider === "hermes-agent") return "hermes-agent";
+  if (provider === "openai" || provider === "anthropic" || provider === "xai") return provider;
+  return null;
+}
+
+export function serializeProviderFields<T extends Record<string, any>>(value: T): T {
+  const next: Record<string, any> = { ...value };
+  if ("initiatedProvider" in next) {
+    next.initiatedProvider = fromPrismaProvider(next.initiatedProvider);
+  }
+  if ("lastUsedProvider" in next) {
+    next.lastUsedProvider = fromPrismaProvider(next.lastUsedProvider);
+  }
+  if ("provider" in next) {
+    next.provider = fromPrismaProvider(next.provider);
+  }
+  if (Array.isArray(next.calls)) {
+    next.calls = next.calls.map((call: Record<string, any>) => serializeProviderFields(call));
+  }
+  return next as T;
+}

server/src/llm/providers.ts

@@ -13,6 +13,18 @@ export function xaiClient() {
   return new OpenAI({ apiKey: env.XAI_API_KEY, baseURL: "https://api.x.ai/v1" });
 }
 
+export function isHermesAgentConfigured() {
+  return Boolean(env.HERMES_AGENT_API_KEY);
+}
+
+export function hermesAgentClient() {
+  if (!env.HERMES_AGENT_API_KEY) throw new Error("HERMES_AGENT_API_KEY not set");
+  return new OpenAI({
+    apiKey: env.HERMES_AGENT_API_KEY,
+    baseURL: env.HERMES_AGENT_API_BASE_URL,
+  });
+}
+
 export function anthropicClient() {
   if (!env.ANTHROPIC_API_KEY) throw new Error("ANTHROPIC_API_KEY not set");
   return new Anthropic({ apiKey: env.ANTHROPIC_API_KEY });

server/src/llm/streaming.ts

@@ -1,13 +1,15 @@
 import { performance } from "node:perf_hooks";
 import { prisma } from "../db.js";
-import { anthropicClient, openaiClient, xaiClient } from "./providers.js";
+import { anthropicClient, hermesAgentClient, openaiClient, xaiClient } from "./providers.js";
 import {
   buildToolLogMessageData,
+  runPlainChatCompletionsStream,
   runToolAwareChatCompletionsStream,
   runToolAwareOpenAIChatStream,
   type ToolExecutionEvent,
 } from "./chat-tools.js";
 import { buildAnthropicConversationMessage, getAnthropicSystemPrompt } from "./message-content.js";
+import { toPrismaProvider } from "./provider-ids.js";
 import type { MultiplexRequest, Provider } from "./types.js";
 
 type StreamUsage = {
@@ -38,7 +40,7 @@ export async function* runMultiplexStream(req: MultiplexRequest): AsyncGenerator
       ? await prisma.llmCall.create({
           data: {
             chatId,
-            provider: req.provider as any,
+            provider: toPrismaProvider(req.provider) as any,
             model: req.model,
             request: req as any,
           },
@@ -51,14 +53,14 @@ export async function* runMultiplexStream(req: MultiplexRequest): AsyncGenerator
       prisma.chat.update({
         where: { id: chatId },
         data: {
-          lastUsedProvider: req.provider as any,
+          lastUsedProvider: toPrismaProvider(req.provider) as any,
           lastUsedModel: req.model,
         },
       }),
       prisma.chat.updateMany({
         where: { id: chatId, initiatedProvider: null },
         data: {
-          initiatedProvider: req.provider as any,
+          initiatedProvider: toPrismaProvider(req.provider) as any,
           initiatedModel: req.model,
         },
       }),
@@ -72,8 +74,8 @@ export async function* runMultiplexStream(req: MultiplexRequest): AsyncGenerator
   let raw: unknown = { streamed: true };
 
   try {
-    if (req.provider === "openai" || req.provider === "xai") {
+    if (req.provider === "openai" || req.provider === "xai" || req.provider === "hermes-agent") {
-      const client = req.provider === "openai" ? openaiClient() : xaiClient();
+      const client = req.provider === "openai" ? openaiClient() : req.provider === "xai" ? xaiClient() : hermesAgentClient();
       const streamEvents =
         req.provider === "openai"
           ? runToolAwareOpenAIChatStream({
@@ -88,6 +90,19 @@ export async function* runMultiplexStream(req: MultiplexRequest): AsyncGenerator
                 chatId: chatId ?? undefined,
               },
             })
+          : req.provider === "hermes-agent"
+            ? runPlainChatCompletionsStream({
+                client,
+                model: req.model,
+                messages: req.messages,
+                temperature: req.temperature,
+                maxTokens: req.maxTokens,
+                logContext: {
+                  provider: req.provider,
+                  model: req.model,
+                  chatId: chatId ?? undefined,
+                },
+              })
           : runToolAwareChatCompletionsStream({
               client,
               model: req.model,

server/src/llm/types.ts

@@ -1,4 +1,6 @@
-export type Provider = "openai" | "anthropic" | "xai";
+export const PROVIDERS = ["openai", "anthropic", "xai", "hermes-agent"] as const;
+
+export type Provider = (typeof PROVIDERS)[number];
 
 export type ChatImageAttachment = {
   kind: "image";

server/src/routes.ts

@@ -10,9 +10,12 @@ import { runMultiplex } from "./llm/multiplexer.js";
 import { runMultiplexStream, type StreamEvent } from "./llm/streaming.js";
 import { getModelCatalogSnapshot } from "./llm/model-catalog.js";
 import { openaiClient } from "./llm/providers.js";
+import { serializeProviderFields, toPrismaProvider } from "./llm/provider-ids.js";
 import { exaClient } from "./search/exa.js";
 import type { ChatAttachment } from "./llm/types.js";
 
+const ProviderSchema = z.enum(["openai", "anthropic", "xai", "hermes-agent"]);
+
 type IncomingChatMessage = {
   role: "system" | "user" | "assistant" | "tool";
   content: string;
@@ -125,7 +128,7 @@ const CompletionStreamBody = z
   .object({
     chatId: z.string().optional(),
     persist: z.boolean().optional(),
-    provider: z.enum(["openai", "anthropic", "xai"]),
+    provider: ProviderSchema,
     model: z.string().min(1),
     messages: z.array(CompletionMessageSchema),
     temperature: z.number().min(0).max(2).optional(),
@@ -591,7 +594,7 @@ export async function registerRoutes(app: FastifyInstance) {
         lastUsedModel: true,
       },
     });
-    return { chats };
+    return { chats: chats.map((chat) => serializeProviderFields(chat)) };
   });
 
   app.post("/v1/chats", async (req) => {
@@ -599,7 +602,7 @@ export async function registerRoutes(app: FastifyInstance) {
     const Body = z
       .object({
         title: z.string().optional(),
-        provider: z.enum(["openai", "anthropic", "xai"]).optional(),
+        provider: ProviderSchema.optional(),
         model: z.string().trim().min(1).optional(),
         messages: z.array(CompletionMessageSchema).optional(),
       })
@@ -625,9 +628,9 @@ export async function registerRoutes(app: FastifyInstance) {
     const chat = await prisma.chat.create({
       data: {
         title: body.title,
-        initiatedProvider: body.provider as any,
+        initiatedProvider: body.provider ? (toPrismaProvider(body.provider) as any) : undefined,
         initiatedModel: body.model,
-        lastUsedProvider: body.provider as any,
+        lastUsedProvider: body.provider ? (toPrismaProvider(body.provider) as any) : undefined,
         lastUsedModel: body.model,
         messages: body.messages?.length
           ? {
@@ -651,7 +654,7 @@ export async function registerRoutes(app: FastifyInstance) {
         lastUsedModel: true,
       },
     });
-    return { chat };
+    return { chat: serializeProviderFields(chat) };
   });
 
   app.patch("/v1/chats/:chatId", async (req) => {
@@ -682,7 +685,7 @@ export async function registerRoutes(app: FastifyInstance) {
       },
     });
     if (!chat) return app.httpErrors.notFound("chat not found");
-    return { chat };
+    return { chat: serializeProviderFields(chat) };
   });
 
   app.post("/v1/chats/title/suggest", async (req) => {
@@ -707,7 +710,7 @@ export async function registerRoutes(app: FastifyInstance) {
       },
     });
     if (!existing) return app.httpErrors.notFound("chat not found");
-    if (existing.title?.trim()) return { chat: existing };
+    if (existing.title?.trim()) return { chat: serializeProviderFields(existing) };
 
     const fallback = body.content.split(/\r?\n/)[0]?.trim().slice(0, 48) || "New chat";
     const suggestedRaw = await generateChatTitle(body.content);
@@ -728,7 +731,7 @@ export async function registerRoutes(app: FastifyInstance) {
       },
     });
 
-    return { chat };
+    return { chat: serializeProviderFields(chat) };
   });
 
   app.delete("/v1/chats/:chatId", async (req) => {
@@ -848,7 +851,7 @@ export async function registerRoutes(app: FastifyInstance) {
       },
     });
 
-    return { chat };
+    return { chat: serializeProviderFields(chat) };
   });
 
   app.post("/v1/searches/:searchId/run", async (req) => {
@@ -994,7 +997,7 @@ export async function registerRoutes(app: FastifyInstance) {
       include: { messages: { orderBy: { createdAt: "asc" } }, calls: { orderBy: { createdAt: "desc" } } },
     });
     if (!chat) return app.httpErrors.notFound("chat not found");
-    return { chat };
+    return { chat: serializeProviderFields(chat) };
   });
 
   app.post("/v1/chats/:chatId/messages", async (req) => {
@@ -1041,7 +1044,7 @@ export async function registerRoutes(app: FastifyInstance) {
 
     const Body = z.object({
       chatId: z.string().optional(),
-      provider: z.enum(["openai", "anthropic", "xai"]),
+      provider: ProviderSchema,
       model: z.string().min(1),
       messages: z.array(CompletionMessageSchema),
       temperature: z.number().min(0).max(2).optional(),

server/tests/chat-tools-streaming.test.ts

@@ -1,6 +1,7 @@
 import assert from "node:assert/strict";
 import test from "node:test";
 import {
+  runPlainChatCompletionsStream,
   runToolAwareChatCompletionsStream,
   runToolAwareOpenAIChatStream,
   type ToolAwareStreamingEvent,
@@ -105,3 +106,37 @@ test("OpenAI-compatible Chat Completions stream emits text deltas as they arrive
   );
   assert.equal(events.at(-1)?.type === "done" ? events.at(-1)?.result.text : null, "Hello");
 });
+
+test("plain Chat Completions stream does not send Sybil-managed tools", async () => {
+  let requestBody: any = null;
+  const client = {
+    chat: {
+      completions: {
+        create: async (body: any) => {
+          requestBody = body;
+          return streamFrom([
+            { choices: [{ delta: { content: "Hi" } }] },
+            { choices: [{ delta: {}, finish_reason: "stop" }] },
+          ]);
+        },
+      },
+    },
+  };
+
+  const events = await collectEvents(
+    runPlainChatCompletionsStream({
+      client: client as any,
+      model: "hermes-agent",
+      messages: [{ role: "user", content: "Say hi" }],
+    })
+  );
+
+  assert.equal(requestBody.model, "hermes-agent");
+  assert.equal(requestBody.stream, true);
+  assert.equal("tools" in requestBody, false);
+  assert.deepEqual(
+    events.map((event) => event.type),
+    ["delta", "done"]
+  );
+  assert.equal(events.at(-1)?.type === "done" ? events.at(-1)?.result.text : null, "Hi");
+});

server/tests/provider-ids.test.ts

@@ -0,0 +1,12 @@
+import assert from "node:assert/strict";
+import test from "node:test";
+import { fromPrismaProvider, serializeProviderFields, toPrismaProvider } from "../src/llm/provider-ids.js";
+
+test("Hermes Agent provider id maps between API and Prisma enum forms", () => {
+  assert.equal(toPrismaProvider("hermes-agent"), "hermes_agent");
+  assert.equal(fromPrismaProvider("hermes_agent"), "hermes-agent");
+  assert.deepEqual(serializeProviderFields({ initiatedProvider: "hermes_agent", lastUsedProvider: "xai" }), {
+    initiatedProvider: "hermes-agent",
+    lastUsedProvider: "xai",
+  });
+});

tui/README.md

@@ -23,7 +23,7 @@ Configuration is environment-only (no in-app settings).
 
 - `SYBIL_TUI_API_BASE_URL`: API base URL. Default: `http://127.0.0.1:8787`
 - `SYBIL_TUI_ADMIN_TOKEN`: optional bearer token for token-mode servers
-- `SYBIL_TUI_DEFAULT_PROVIDER`: `openai` | `anthropic` | `xai` (default: `openai`)
+- `SYBIL_TUI_DEFAULT_PROVIDER`: `openai` | `anthropic` | `xai` | `hermes-agent` (default: `openai`)
 - `SYBIL_TUI_DEFAULT_MODEL`: optional default model name
 - `SYBIL_TUI_SEARCH_NUM_RESULTS`: results per search run (default: `10`)
 

tui/src/config.ts

@@ -1,6 +1,6 @@
 import type { Provider } from "./types.js";
 
-const PROVIDERS: Provider[] = ["openai", "anthropic", "xai"];
+const PROVIDERS: Provider[] = ["openai", "anthropic", "xai", "hermes-agent"];
 
 function normalizeBaseUrl(value: string) {
   const trimmed = value.trim();

tui/src/index.ts

@@ -39,11 +39,13 @@ type ToolLogMetadata = {
   resultPreview?: string | null;
 };
 
-const PROVIDERS: Provider[] = ["openai", "anthropic", "xai"];
+const BASE_PROVIDERS: Provider[] = ["openai", "anthropic", "xai"];
+const PROVIDERS: Provider[] = [...BASE_PROVIDERS, "hermes-agent"];
 const PROVIDER_FALLBACK_MODELS: Record<Provider, string[]> = {
   openai: ["gpt-4.1-mini"],
   anthropic: ["claude-3-5-sonnet-latest"],
   xai: ["grok-3-mini"],
+  "hermes-agent": ["hermes-agent"],
 };
 
 const EMPTY_MODEL_CATALOG: ModelCatalogResponse["providers"] = {
@@ -74,6 +76,7 @@ function getProviderLabel(provider: Provider | null | undefined) {
   if (provider === "openai") return "OpenAI";
   if (provider === "anthropic") return "Anthropic";
   if (provider === "xai") return "xAI";
+  if (provider === "hermes-agent") return "Hermes Agent";
   return "";
 }
 
@@ -159,6 +162,10 @@ function getModelOptions(catalog: ModelCatalogResponse["providers"], provider: P
   return PROVIDER_FALLBACK_MODELS[provider];
 }
 
+function getVisibleProviders(catalog: ModelCatalogResponse["providers"]) {
+  return PROVIDERS.filter((provider) => provider !== "hermes-agent" || catalog[provider] !== undefined);
+}
+
 function pickProviderModel(options: string[], preferred: string | null, fallback: string | null = null) {
   if (fallback && options.includes(fallback)) return fallback;
   if (preferred && options.includes(preferred)) return preferred;
@@ -202,6 +209,7 @@ async function main() {
     openai: null,
     anthropic: null,
     xai: null,
+    "hermes-agent": null,
   };
   let model: string = config.defaultModel ?? pickProviderModel(getModelOptions(modelCatalog, provider), null);
   let errorMessage: string | null = null;
@@ -1257,8 +1265,10 @@ async function main() {
   }
 
   function cycleProvider() {
-    const currentIndex = PROVIDERS.indexOf(provider);
+    const visibleProviders = getVisibleProviders(modelCatalog);
-    const nextProvider: Provider = PROVIDERS[(currentIndex + 1) % PROVIDERS.length] ?? "openai";
+    const cycleProviders = visibleProviders.length ? visibleProviders : BASE_PROVIDERS;
+    const currentIndex = Math.max(0, cycleProviders.indexOf(provider));
+    const nextProvider: Provider = cycleProviders[(currentIndex + 1) % cycleProviders.length] ?? "openai";
     provider = nextProvider;
     syncModelForProvider();
     updateUI();

tui/src/types.ts

@@ -1,4 +1,4 @@
-export type Provider = "openai" | "anthropic" | "xai";
+export type Provider = "openai" | "anthropic" | "xai" | "hermes-agent";
 
 export type ProviderModelInfo = {
   models: string[];
@@ -7,7 +7,7 @@ export type ProviderModelInfo = {
 };
 
 export type ModelCatalogResponse = {
-  providers: Record<Provider, ProviderModelInfo>;
+  providers: Partial<Record<Provider, ProviderModelInfo>>;
 };
 
 export type ChatSummary = {

web/src/App.tsx

@@ -95,6 +95,7 @@ const PROVIDER_FALLBACK_MODELS: Record<Provider, string[]> = {
   openai: ["gpt-4.1-mini"],
   anthropic: ["claude-3-5-sonnet-latest"],
   xai: ["grok-3-mini"],
+  "hermes-agent": ["hermes-agent"],
 };
 
 const EMPTY_MODEL_CATALOG: ModelCatalogResponse["providers"] = {
@@ -103,6 +104,9 @@ const EMPTY_MODEL_CATALOG: ModelCatalogResponse["providers"] = {
   xai: { models: [], loadedAt: null, error: null },
 };
 
+const BASE_PROVIDERS: Provider[] = ["openai", "anthropic", "xai"];
+const ALL_PROVIDERS: Provider[] = [...BASE_PROVIDERS, "hermes-agent"];
+
 const MODEL_PREFERENCES_STORAGE_KEY = "sybil:modelPreferencesByProvider";
 const QUICK_QUESTION_MODEL_SELECTION_STORAGE_KEY = "sybil:quickQuestionModelSelection";
 
@@ -117,6 +121,7 @@ const EMPTY_MODEL_PREFERENCES: ProviderModelPreferences = {
   openai: null,
   anthropic: null,
   xai: null,
+  "hermes-agent": null,
 };
 const EMPTY_ACTIVE_RUNS: ActiveRunsState = {
   chats: {},
@@ -193,6 +198,10 @@ function getModelOptions(catalog: ModelCatalogResponse["providers"], provider: P
   return PROVIDER_FALLBACK_MODELS[provider];
 }
 
+function getVisibleProviders(catalog: ModelCatalogResponse["providers"]) {
+  return ALL_PROVIDERS.filter((provider) => provider !== "hermes-agent" || catalog[provider] !== undefined);
+}
+
 function getReplyScrollBufferHeight() {
   if (typeof window === "undefined") return REPLY_SCROLL_BUFFER_MIN;
   return Math.min(
@@ -308,6 +317,8 @@ function loadStoredModelPreferences() {
       openai: typeof parsed.openai === "string" && parsed.openai.trim() ? parsed.openai.trim() : null,
       anthropic: typeof parsed.anthropic === "string" && parsed.anthropic.trim() ? parsed.anthropic.trim() : null,
       xai: typeof parsed.xai === "string" && parsed.xai.trim() ? parsed.xai.trim() : null,
+      "hermes-agent":
+        typeof parsed["hermes-agent"] === "string" && parsed["hermes-agent"].trim() ? parsed["hermes-agent"].trim() : null,
     };
   } catch {
     return EMPTY_MODEL_PREFERENCES;
@@ -315,7 +326,7 @@ function loadStoredModelPreferences() {
 }
 
 function normalizeStoredProvider(value: unknown): Provider {
-  return value === "anthropic" || value === "xai" || value === "openai" ? value : "openai";
+  return value === "anthropic" || value === "xai" || value === "openai" || value === "hermes-agent" ? value : "openai";
 }
 
 function normalizeStoredModelPreferences(value: unknown): ProviderModelPreferences {
@@ -325,6 +336,8 @@ function normalizeStoredModelPreferences(value: unknown): ProviderModelPreferenc
       openai: typeof parsed.openai === "string" && parsed.openai.trim() ? parsed.openai.trim() : null,
       anthropic: typeof parsed.anthropic === "string" && parsed.anthropic.trim() ? parsed.anthropic.trim() : null,
       xai: typeof parsed.xai === "string" && parsed.xai.trim() ? parsed.xai.trim() : null,
+      "hermes-agent":
+        typeof parsed["hermes-agent"] === "string" && parsed["hermes-agent"].trim() ? parsed["hermes-agent"].trim() : null,
     };
 }
 
@@ -354,6 +367,7 @@ function getProviderLabel(provider: Provider | null | undefined) {
   if (provider === "openai") return "OpenAI";
   if (provider === "anthropic") return "Anthropic";
   if (provider === "xai") return "xAI";
+  if (provider === "hermes-agent") return "Hermes Agent";
   return "";
 }
 
@@ -715,6 +729,8 @@ export default function App() {
   const wasSendingRef = useRef(false);
   const pendingReplyScrollRef = useRef(false);
   const transcriptTailSpacerHeightRef = useRef(TRANSCRIPT_BOTTOM_GAP);
+  const transcriptTailSpacerSettleFrameRef = useRef<number | null>(null);
+  const transcriptViewKeyRef = useRef<string | null>(null);
   const [contextMenu, setContextMenu] = useState<ContextMenuState | null>(null);
   const [isMobileSidebarOpen, setIsMobileSidebarOpen] = useState(false);
   const [sidebarQuery, setSidebarQuery] = useState("");
@@ -739,6 +755,7 @@ export default function App() {
   const settleTranscriptTailSpacer = () => {
     const container = transcriptContainerRef.current;
     const currentSpacerHeight = transcriptTailSpacerHeightRef.current;
+    if (currentSpacerHeight <= TRANSCRIPT_BOTTOM_GAP) return;
     if (!container) {
       setTranscriptTailSpacer(TRANSCRIPT_BOTTOM_GAP);
       return;
@@ -746,7 +763,22 @@ export default function App() {
 
     const scrollHeightWithoutSpacer = container.scrollHeight - currentSpacerHeight;
     const requiredSpacerHeight = container.scrollTop + container.clientHeight - scrollHeightWithoutSpacer;
-    setTranscriptTailSpacer(requiredSpacerHeight);
+    setTranscriptTailSpacer(Math.min(currentSpacerHeight, requiredSpacerHeight));
+  };
+
+  const requestSettleTranscriptTailSpacer = () => {
+    if (transcriptTailSpacerHeightRef.current <= TRANSCRIPT_BOTTOM_GAP) return;
+    if (typeof window === "undefined") {
+      settleTranscriptTailSpacer();
+      return;
+    }
+    if (transcriptTailSpacerSettleFrameRef.current !== null) {
+      window.cancelAnimationFrame(transcriptTailSpacerSettleFrameRef.current);
+    }
+    transcriptTailSpacerSettleFrameRef.current = window.requestAnimationFrame(() => {
+      transcriptTailSpacerSettleFrameRef.current = null;
+      settleTranscriptTailSpacer();
+    });
   };
 
   const focusComposer = () => {
@@ -963,6 +995,7 @@ export default function App() {
 
   const providerModelOptions = useMemo(() => getModelOptions(modelCatalog, provider), [modelCatalog, provider]);
   const quickProviderModelOptions = useMemo(() => getModelOptions(modelCatalog, quickProvider), [modelCatalog, quickProvider]);
+  const providerOptions = useMemo(() => getVisibleProviders(modelCatalog), [modelCatalog]);
 
   useEffect(() => {
     if (model.trim()) return;
@@ -1017,6 +1050,7 @@ export default function App() {
   }, [quickPrompt, isQuickQuestionOpen]);
 
   const selectedKey = selectedItem ? `${selectedItem.kind}:${selectedItem.id}` : null;
+  const transcriptViewKey = draftKind ? `draft:${draftKind}` : selectedKey ?? "empty";
   const selectedChatPendingState = selectedItem?.kind === "chat" ? pendingChatStates[selectedItem.id] ?? null : null;
   const selectedSearchRunState = selectedItem?.kind === "search" ? runningSearchStates[selectedItem.id] ?? null : null;
   const selectedChatIsActive = selectedItem?.kind === "chat" && (!!selectedChatPendingState || !!activeRuns.chats[selectedItem.id]);
@@ -1038,11 +1072,13 @@ export default function App() {
   };
 
   useEffect(() => {
+    const didViewChange = transcriptViewKeyRef.current !== transcriptViewKey;
+    transcriptViewKeyRef.current = transcriptViewKey;
     shouldAutoScrollRef.current = true;
-    if (!isSendingActiveChat) {
+    if (didViewChange && !pendingReplyScrollRef.current) {
       setTranscriptTailSpacer(TRANSCRIPT_BOTTOM_GAP);
     }
-  }, [isSendingActiveChat, selectedKey]);
+  }, [transcriptViewKey]);
 
   useEffect(() => {
     selectedItemRef.current = selectedItem;
@@ -1101,6 +1137,10 @@ export default function App() {
 
   useEffect(() => {
     return () => {
+      if (transcriptTailSpacerSettleFrameRef.current !== null) {
+        window.cancelAnimationFrame(transcriptTailSpacerSettleFrameRef.current);
+        transcriptTailSpacerSettleFrameRef.current = null;
+      }
       for (const controller of chatStreamAbortRefs.current.values()) {
         controller.abort();
       }
@@ -1129,6 +1169,16 @@ export default function App() {
     if (selectedChatPendingState) return selectedChatPendingState.messages.filter(isDisplayableMessage);
     return messages.filter(isDisplayableMessage);
   }, [messages, selectedChatPendingState]);
+  const displayMessagesLayoutKey = useMemo(
+    () => displayMessages.map((message) => `${message.id}:${message.content.length}`).join("|"),
+    [displayMessages]
+  );
+
+  useEffect(() => {
+    if (isSearchMode || isSendingActiveChat) return;
+    requestSettleTranscriptTailSpacer();
+  }, [displayMessagesLayoutKey, isSearchMode, isSendingActiveChat, selectedKey]);
+
   const quickAnswerText = useMemo(() => {
     for (let index = quickQuestionMessages.length - 1; index >= 0; index -= 1) {
       const message = quickQuestionMessages[index];
@@ -1465,6 +1515,11 @@ export default function App() {
   };
 
   const handleSendChat = async (content: string, attachments: ChatAttachment[]): Promise<SidebarSelection> => {
+    const selectedModel = model.trim();
+    if (!selectedModel) {
+      throw new Error("No model available for selected provider");
+    }
+
     pendingReplyScrollRef.current = true;
     expandTranscriptTailSpacer(getReplyScrollBufferHeight());
 
@@ -1548,11 +1603,6 @@ export default function App() {
       },
     ];
 
-    const selectedModel = model.trim();
-    if (!selectedModel) {
-      throw new Error("No model available for selected provider");
-    }
-
     const chatSummary = chats.find((chat) => chat.id === chatId);
     const hasExistingTitle = Boolean(selectedChat?.id === chatId ? selectedChat.title?.trim() : chatSummary?.title?.trim());
     if (!hasExistingTitle && !pendingTitleGenerationRef.current.has(chatId)) {
@@ -1672,13 +1722,17 @@ export default function App() {
       if (currentSelection?.kind === "chat" && currentSelection.id === chatId) {
         await refreshChat(chatId);
       }
-      settleTranscriptTailSpacer();
       removePendingChatState(chatId);
       removeActiveRun("chat", chatId);
+      if (currentSelection?.kind === "chat" && currentSelection.id === chatId) {
+        requestSettleTranscriptTailSpacer();
+      }
       return { kind: "chat", id: chatId };
     } catch (err) {
       removePendingChatState(chatId);
       removeActiveRun("chat", chatId);
+      pendingReplyScrollRef.current = false;
+      setTranscriptTailSpacer(TRANSCRIPT_BOTTOM_GAP);
       throw err;
     }
   };
@@ -1939,6 +1993,9 @@ export default function App() {
       chatStreamAbortRefs.current.delete(chatId);
       removePendingChatState(chatId);
       removeActiveRun("chat", chatId);
+      if (isCurrentSelection(target)) {
+        requestSettleTranscriptTailSpacer();
+      }
     }
   };
 
@@ -2288,6 +2345,10 @@ export default function App() {
         sentTarget = await handleSendChat(content, attachments);
       }
     } catch (err) {
+      if (!sentAsSearch) {
+        pendingReplyScrollRef.current = false;
+        setTranscriptTailSpacer(TRANSCRIPT_BOTTOM_GAP);
+      }
       const message = err instanceof Error ? err.message : String(err);
       if (message.includes("bearer token")) {
         handleAuthFailure(message);
@@ -2512,9 +2573,11 @@ export default function App() {
                     }}
                     disabled={isActiveSelectionSending}
                   >
-                    <option value="openai">OpenAI</option>
+                    {providerOptions.map((candidate) => (
-                    <option value="anthropic">Anthropic</option>
+                      <option key={candidate} value={candidate}>
-                    <option value="xai">xAI</option>
+                        {getProviderLabel(candidate)}
+                      </option>
+                    ))}
                   </select>
                   <ModelCombobox
                     options={providerModelOptions}
@@ -2547,6 +2610,9 @@ export default function App() {
               if (!container) return;
               const distanceFromBottom = container.scrollHeight - container.scrollTop - container.clientHeight;
               shouldAutoScrollRef.current = distanceFromBottom < 96;
+              if (!isSearchMode && !isSendingActiveChat && distanceFromBottom > 0) {
+                settleTranscriptTailSpacer();
+              }
             }}
           >
             {!isSearchMode ? (
@@ -2758,9 +2824,11 @@ export default function App() {
                   disabled={isQuickQuestionSending || isConvertingQuickQuestion}
                   aria-label="Quick question provider"
                 >
-                  <option value="openai">OpenAI</option>
+                  {providerOptions.map((candidate) => (
-                  <option value="anthropic">Anthropic</option>
+                    <option key={candidate} value={candidate}>
-                  <option value="xai">xAI</option>
+                      {getProviderLabel(candidate)}
+                    </option>
+                  ))}
                 </select>
                 <ModelCombobox
                   options={quickProviderModelOptions}

web/src/index.css

@@ -206,17 +206,31 @@ textarea {
 }
 
 .md-content code {
-  background: hsl(288 22% 23%);
+  background: hsl(249 40% 10% / 0.78);
-  border-radius: 0.25rem;
+  border-radius: 0.3rem;
   padding: 0.05rem 0.3rem;
   font-size: 0.86em;
+  box-decoration-break: clone;
+  -webkit-box-decoration-break: clone;
 }
 
 .md-content pre {
   overflow-x: auto;
-  border-radius: 0.5rem;
+  border: 1px solid hsl(253 31% 29% / 0.72);
-  background: hsl(287 28% 13%);
+  border-radius: 0.625rem;
-  padding: 0.6rem 0.75rem;
+  background: hsl(249 40% 10% / 0.82);
+  padding: 0.75rem;
+  box-shadow: inset 0 1px 0 hsl(258 80% 88% / 0.05);
+}
+
+.md-content pre code {
+  display: block;
+  background: transparent;
+  border-radius: 0;
+  padding: 0;
+  font-size: 0.88em;
+  line-height: 1.55;
+  white-space: pre;
 }
 
 .md-content a {

web/src/lib/api.ts

@@ -127,7 +127,7 @@ export type CompletionRequestMessage = {
   attachments?: ChatAttachment[];
 };
 
-export type Provider = "openai" | "anthropic" | "xai";
+export type Provider = "openai" | "anthropic" | "xai" | "hermes-agent";
 
 export type ProviderModelInfo = {
   models: string[];
@@ -136,7 +136,7 @@ export type ProviderModelInfo = {
 };
 
 export type ModelCatalogResponse = {
-  providers: Record<Provider, ProviderModelInfo>;
+  providers: Partial<Record<Provider, ProviderModelInfo>>;
 };
 
 export type ActiveRunsResponse = {