* feat: add opencode support
* fix: stabilize opencode session startup
* fix: /models
* fix: improveUI for commands
* fix: format commands.js
* feat: load models through provider adapters
Provider model selection had outgrown a single hardcoded service.
The old service mixed shared caching with provider catalogs and CLI lookup details.
That made stale model lists more likely as providers changed on separate schedules.
Move model discovery behind each provider so lookup lives next to the integration.
The shared service now focuses on provider resolution, caching, persistence, and dedupe.
Return cache metadata and add bypassCache because model availability changes outside the app.
The UI and /models command can show freshness and let users force a provider refresh.
Surface model descriptions while keeping fallback catalogs for unavailable CLIs or SDKs.
* feat(models): resolve active session models through provider adapters
The model inventory command was showing a mix of catalog defaults and
composer-local state instead of the model that is actually active for a
real provider session. That made /models, /cost, and /status
misleading once a session had already started, especially for providers
whose effective runtime model can differ from the optimistic model value
held in the UI.
Introduce an explicit getCurrentActiveModel() contract on
IProviderModels so model resolution lives next to each provider's
catalog logic and uses the provider-native source of truth:
- Claude reads the init event from a resumed stream-json run
- Codex reads model from ~/.codex/config.toml
- Cursor reads lastUsedModel from the chat store.db
- OpenCode reads the persisted session model from opencode.db
- Gemini intentionally returns its default because the CLI does not
provide a reliable active-session lookup
Keep the returned shape intentionally minimal ({ model }). The goal is
to expose only what downstream command consumers need and avoid leaking
provider-specific metadata into a shared transport shape that would
create extra UI coupling and future cleanup cost.
Also make command behavior session-aware: when there is no concrete
session id, do not spawn provider processes or inspect provider session
storage just to answer /models, /cost, or /status. In a new-session
view the correct answer is simply the provider default, and doing more
work there adds latency and unnecessary side effects for no user value.
As part of this, centralize two supporting concerns:
- add a shared helper for building the default current-model result from
a provider catalog so fallbacks stay aligned with DEFAULT
- move leaf-directory validation into shared utils so Cursor session
readers and model lookup code enforce the same path-safety rule
Tests were expanded to cover both the new service delegation path and
the sessionless command behavior, while keeping cache-sensitive tests
isolated from persisted host cache state.
Why this change:
- command output should reflect the model actually driving a session
- new-session views should stay fast and side-effect free
- provider-specific active-model lookup should not be scattered across
routes or UI code
- fallback behavior should be explicit, consistent, and limited to the
provider default when no true active model can be resolved
* feat: support session-scoped model overrides
Model selection was acting like a provider-level preference.
That made resumed sessions drift back to a default or request-time model.
Users expect /models changes made inside a conversation to affect that session.
Store explicit session choices in app-owned ~/.cloudcli state.
This avoids editing provider transcripts or native provider config.
Resolve the effective model before launching each provider runtime.
Claude, Cursor, Codex, Gemini, and OpenCode now honor stored resume choices.
Expose a backend active-model change endpoint for existing sessions.
The models modal can now distinguish default changes from session overrides.
It also shows when a selected model will apply on the next response.
For Claude, stop probing active model state by resuming with a dummy prompt.
Read the indexed JSONL transcript from the end instead.
This preserves provider history while honoring /model stdout or model fields.
Add service tests for adapter delegation and resume-model precedence.
The tests keep cache state, override state, and requested fallback separate.
* feat: make command modal more compact
* fix: preserve opencode session creation events
OpenCode emits the real session id asynchronously on its first JSON output. The runner
registered that id from a helper that could not see the spawned process because
the process reference was scoped inside the model-resolution callback. That
ReferenceError was swallowed by the generic JSON parse fallback, so the client
never received session_created. Without that event, a new OpenCode chat stayed
on / and the assistant stream was not attached to the new session view.
Keep the process reference in the outer spawn scope so registration can update
the active-process map and websocket writer as soon as OpenCode announces the
session id. Split JSON parsing from event processing so malformed non-JSON
output can still stream as raw text, while registration or adapter failures are
surfaced as real errors instead of being hidden as assistant content.
Add a fake opencode executable regression test to lock in the expected lifecycle
ordering: session_created must be sent before live assistant messages, and the
same session id must carry through stream_end and complete.
* fix: clarify model refresh and onboarding providers
OpenCode is now a supported chat provider, but first-run onboarding still only offered
Claude, Cursor, Codex, and Gemini. That made OpenCode harder to discover and
forced users to finish setup before finding the provider in settings or chat.
Adding it to onboarding keeps first-run setup aligned with the providers the
application already supports elsewhere.
The model refresh control was also doing too much visual work. In the new chat
model picker, the previous Hard Refresh label looked like the dialog heading,
which made the primary task unclear. Users open that dialog to choose a model;
refreshing catalogs is only a secondary maintenance action for stale cached
provider model lists.
Rename and reposition the refresh affordance so the model picker reads as a
model picker first. The copy now explains why catalogs are cached, when a refresh
is useful, and that the refresh checks every provider. The /models modal gets the
same clarification so both model-selection surfaces describe the cache behavior
consistently.
* fix: format opencode model catalog labels
OpenCode returns provider-prefixed ids directly from the CLI. Passing those ids through as
labels made the model picker hard to scan: users saw values like
anthropic/claude-3-5-sonnet-20241022 or lowercased, hyphen-split text instead
of readable model names.
Keep the exact OpenCode id as the option value because that is what the CLI
expects, but derive a presentation label for the frontend. The formatter is
intentionally generic rather than a catalog of known providers. It handles common
identifier structure such as provider/model, hyphen-delimited words, v-prefixed
versions, adjacent numeric version tokens, and 8-digit date suffixes.
This keeps OpenCode usable as its model list expands across many upstream
providers without requiring code changes for every new provider or model family.
The description keeps the raw provider-prefixed id visible so users can still
confirm the precise model being selected.
* feat: add more fallback models for cursor
* docs: move model catalog out of shared
The model catalog is no longer a frontend/backend runtime contract.
Keeping it under shared made ownership misleading. It implied the catalog was
application code shared by runtime consumers, even though it now only supports
README links and public API documentation.
Move the catalog into public so it lives beside the docs surfaces that need it.
This gives the API docs a stable, served module and gives README readers a
linkable source without suggesting frontend or backend runtime dependency.
Render the API docs model list from the exported provider registry instead of a
hardcoded Claude/Cursor/Codex subset. That keeps Gemini and OpenCode visible and
makes future provider documentation changes flow through one docs-specific file.
Update README links, provider maintenance notes, and package files so published
artifacts include the standalone docs page and model catalog without relying on
the old shared path.
* fix: simplify empty-state model selector
Keep the provider empty state focused on the setup action users need there:
choosing a model.
The refresh control, cache timestamp, and refresh explanation made the dialog feel
like a cache-management surface.
That extra action is out of place in the empty state, where the goal is to start
a chat with the selected provider and model.
Remove the refresh-specific UI from ProviderSelectionEmptyState and drop the
now-unused refresh/cache props from the ChatMessagesPane pass-through.
Refresh behavior remains available in the dedicated command result flow.
* feat(i18n): add Turkish (tr) language support
Add comprehensive Turkish localization for the UI, following the
existing i18n pattern established by Japanese (#384), Russian (#514),
and German (#525) language support.
Changes:
- Add Turkish translation files for all 7 namespaces
(auth, chat, codeEditor, common, settings, sidebar, tasks)
- Register Turkish locale in config.js with all resources
- Add Turkish entry to languages.js (value: tr, nativeName: Türkçe)
- Update .gitignore to allow src/i18n/locales/tr/tasks.json
(matches existing en/ja/ru/de exceptions)
Translation details:
- 934 total strings translated (100% coverage, matches en.json key count)
- Translated by a native Turkish speaker with software engineering
background; terminology reviewed against conventional Turkish
tech community usage.
- Technical terms kept in English per Turkish dev community norms:
Claude, Cursor, Codex, Gemini, CLI, MCP, PRD, JSON, YAML, stdio,
http, commit, branch, token, prompt, minimap, sandbox, YOLO.
- Informal second-person singular (\"sen\") used throughout — fits the
developer-facing nature of the UI.
- All interpolation placeholders preserved exactly (e.g. {{count}},
{{projectName}}, {{email}}).
- i18next plural keys (_one/_other) kept intact.
Verification:
- Key structure parity with en.json confirmed (jq paths diff empty)
- All 38 unique interpolation variables preserved
- npm run build passes cleanly
* docs(readme): add Turkish README and language switcher links
Add README.tr.md — full Turkish translation of the main README,
following the structure of existing README.de.md / README.ja.md /
README.ko.md / README.ru.md / README.zh-CN.md.
Update the language switcher row in all 6 existing README variants
to include a Turkish link (matches the pattern used by #534 for the
German language link addition).
---------
Co-authored-by: Simos Mikelatos <simosmik@gmail.com>
* feat: add WebSocket proxy for plugin backends
Adds /plugin-ws/:name route that proxies authenticated WebSocket
connections to plugin server subprocesses, enabling real-time
bidirectional communication for plugins like web-terminal.
* chore: update README with the plugin
* docs: standardize hero badges and language order across README translations
* fix: label README command blocks as bash
* docs: translate remaining sections to Japanese
* fix(readme): remove sponsor duplicate in Japanese readme
* fix(readme): japanese translation fixes
* fix(readme): remove duplicate sections
---------
Co-authored-by: Haile <118998054+blackmammoth@users.noreply.github.com>
* feat: add Russian locale
- Add ru translations and register namespaces
- Add Russian to supported languages list
- Ignore .gemini workspace config
* fix: improve Russian plural forms in sidebar translations
Add proper Russian plural forms (few/many) for correct grammar with different count values
* docs(readme): add Russian translation and fix language switcher order
- Create README.ru.md based on the current English README.
- Update language switchers in all localized README files so
English comes first, Russian second, and the remaining
languages follow.
- Fix the issue where the current language was not shown
correctly in the switcher for some localized README files
* fix(readme): fix language switcher positions and markdown issues
- Fix language switcher positions in README.md.
- Add bash language tags to command code blocks in README.ru.md.
* fix(readme): fix tool setup step numbering
- Fix tool setup step numbering in README.md and localized README files.
* fix(gitignore): allow translation task files to be tracked
Add exceptions to .gitignore for task translation files across multiple locales
(en, ja, ru, ko, zh-CN) to enable version control of translated content while
keeping generated task files ignored.
Co-Authored-By: Claude Haiku 4.5 <noreply@anthropic.com>
* feat(i18n): add Russian translation for tasks
Add Russian locale translation file for TaskMaster task management interface.
Co-Authored-By: Claude Haiku 4.5 <noreply@anthropic.com>
* fix: ignore missing tasks.json files for ko and zh-cn locales
* Delete .worktrees directory
---------
Co-authored-by: Claude Haiku 4.5 <noreply@anthropic.com>
* feat: new plugin system
* Potential fix for code scanning alert no. 312: Uncontrolled data used in path expression
Co-authored-by: Copilot Autofix powered by AI <62310815+github-advanced-security[bot]@users.noreply.github.com>
* Update manifest.json
* feat(plugins): add SVG icon support with authenticated inline rendering
* fix: coderabbit changes and new plugin name & repo
* fix: design changes to plugins settings tab
* fix(plugins): prevent git arg injection, add repo URL detection
* fix: lint errors and deleting plugin error on windows
* fix: coderabbit nitpick comments
* fix(plugins): harden path traversal and respect enabled state
Use realpathSync to canonicalize paths before the plugin asset
boundary check, preventing symlink-based traversal bypasses that
could escape the plugin directory.
PluginTabContent now guards on plugin.enabled before mounting the
plugin module, and re-mounts when the enabled state changes so
toggling a plugin takes effect without a page reload.
PluginIcon safely handles a missing iconFile prop and skips
processing non-OK fetch responses instead of attempting to parse
error bodies as SVG.
Register 'plugins' as a known main tab so the settings router
preserves the tab on navigation.
* fix(plugins): support concurrent plugin updates
Replace single updatingPlugin string state with a Set to allow
multiple plugins to update simultaneously. Also disable the update
button and show a descriptive tooltip when a plugin has no git
remote configured.
* fix(plugins): async shutdown and asset/RPC fixes
Await stopPluginServer/stopAllPlugins in signal handlers and route
handlers so process exit and state transitions wait for clean plugin
shutdown instead of racing ahead.
Validate asset paths are regular files before streaming to prevent
directory traversal returning unexpected content; add a stream error
handler to avoid unhandled crashes on read failures.
Fix RPC proxy body detection to use the content-length header instead
of Object.keys, so falsy but valid JSON payloads (null, false, 0, {})
are forwarded correctly to plugin servers.
Track in-flight start operations via a startingPlugins map to prevent
duplicate concurrent plugin starts.
* refactor(git-panel): simplify setCommitMessage with plain function
* fix(plugins): harden input validation and scan reliability
- Validate plugin names against [a-zA-Z0-9_-] allowlist in
manifest and asset routes to prevent path traversal via URL
- Strip embedded credentials (user:pass@) from git remote URLs
before exposing them to the client
- Skip .tmp-* directories during scan to avoid partial installs
from in-progress updates appearing as broken plugins
- Deduplicate plugins sharing the same manifest name to prevent
ambiguous state
- Guard RPC proxy error handler against writing to an already-sent
response, preventing uncaught exceptions on aborted requests
* fix(git-panel): reset changes view on project switch
* refactor: move plugin content to /view folder
* fix: resolve type error in MobileNav and PluginTabContent components
---------
Co-authored-by: Copilot Autofix powered by AI <62310815+github-advanced-security[bot]@users.noreply.github.com>
Co-authored-by: Haileyesus <118998054+blackmammoth@users.noreply.github.com>
Co-authored-by: Haileyesus <something@gmail.com>
* feat: improve version comparison logic in useVersionCheck hook
* refactor: useVersionCheck.js to typescript
* refactor: move VersionUpgradeModal component to its own file and remove it from AppContent component
* refactor: improve VersionUpgradeModal props and extract ReleaseInfo type
Using useVersionCheck hook in 2 places caused github requests to be made twice, which is not ideal.
* refactor: handleUpdateNow function with useCallback and error display in VersionUpgradeModal
* refactor: move isOpen check to the correct position in VersionUpgradeModal
* refactor: move VersionUpgradeModal and collapsed sidebar to Sidebar component from App.jsx
* refactor: remove unused SettingsIcon import from App.jsx
* refactor: move formatTimeAgo function to dateUtils.ts
* refactor: replace useLocalStorage with useUiPreferences for better state management in AppContent
* refactor: use shared props for Sidebar props in AppContent
* refactor: remove showQuickSettings state and toggle from AppContent, manage isOpen state directly in QuickSettingsPanel
* refactor: move preference props directly to QuickSettingsPanel and MainContent
* refactor: remove unused isPWA prop
* refactor: remove unused isPWA prop from AppContent
* refactor: remove unused generatingSummary state from Sidebar component
* refactor: remove unused isPWA prop from MainContent component
* refactor: use usePrefrences for sidebar visibility in Sidebar component
* refactor: extract device detection into hook and localize PWA handling to Sidebar
- Add new `useDeviceSettings` hook (`src/hooks/useDeviceSettings.ts`) to centralize
device-related state:
- exposes `isMobile` and `isPWA`
- supports options: `mobileBreakpoint`, `trackMobile`, `trackPWA`
- listens to window resize for mobile updates
- listens to `display-mode: standalone` changes for PWA updates
- includes `matchMedia.addListener/removeListener` fallback for older environments
- Update `AppContent` (`src/App.jsx`) to consume `isMobile` from
`useDeviceSettings({ trackPWA: false })`:
- remove local `isMobile` state/effect
- remove local `isPWA` state/effect
- keep existing `isMobile` behavior for layout and mobile sidebar flow
- stop passing `isPWA` into `Sidebar` props
- Update `Sidebar` (`src/components/Sidebar.jsx`) to own PWA detection:
- consume `isPWA` from `useDeviceSettings({ trackMobile: false })`
- add effect to toggle `pwa-mode` class on `document.documentElement` and `document.body`
- retain use of `isMobile` prop from `App` for sidebar/mobile rendering decisions
Why:
- removes duplicated device-detection logic from `AppContent`
- makes device-state logic reusable and easier to maintain
- keeps PWA-specific behavior where it is actually used (`Sidebar`)
* chore(to-remove): comment todo's
* refactor: remove unused createNewProject and cancelNewProject functions from Sidebar component
* refactor(sidebar): extract typed app/sidebar architecture and split Sidebar into modular components
- Replace `src/App.jsx` with `src/App.tsx` and move route-level UI orchestration into `src/components/app/AppContent.tsx`.
This separates provider/bootstrap concerns from runtime app layout logic, keeps route definitions minimal, and improves readability of the root app entry.
- Introduce `src/hooks/useProjectsState.ts` to centralize project/session/sidebar state management previously embedded in `App.jsx`.
This keeps the existing behavior for:
project loading,
Cursor session hydration,
WebSocket `loading_progress` handling,
additive-update protection for active sessions,
URL-based session selection,
sidebar refresh/delete/new-session flows.
The hook now exposes a typed `sidebarSharedProps` contract and typed handlers used by `AppContent`.
- Introduce `src/hooks/useSessionProtection.ts` for active/processing session lifecycle logic.
This preserves session-protection behavior while isolating `activeSessions`, `processingSessions`, and temporary-session replacement into a dedicated reusable hook.
- Replace monolithic `src/components/Sidebar.jsx` with typed `src/components/Sidebar.tsx` as a thin orchestrator.
`Sidebar.tsx` now focuses on wiring controller state/actions, modal visibility, collapsed mode, and version modal behavior instead of rendering every UI branch inline.
- Add `src/hooks/useSidebarController.ts` to encapsulate sidebar interaction/state logic.
This includes expand/collapse state, inline project/session editing state, project starring/sorting/filtering, lazy session pagination, delete confirmations, rename/delete actions, refresh state, and mobile touch click handling.
- Add strongly typed sidebar domain models in `src/components/sidebar/types.ts` and move sidebar-derived helpers into `src/components/sidebar/utils.ts`.
Utility coverage now includes:
session provider normalization,
session view-model creation (name/time/activity/message count),
project sorting/filtering,
task indicator status derivation,
starred-project persistence and readbacks.
- Split sidebar rendering into focused components under `src/components/sidebar/`:
`SidebarContent.tsx` for top-level sidebar layout composition.
`SidebarProjectList.tsx` for project-state branching and project iteration.
`SidebarProjectsState.tsx` for loading/empty/no-search-result placeholders.
`SidebarProjectItem.tsx` for per-project desktop/mobile header rendering and actions.
`SidebarProjectSessions.tsx` for expanded session area, skeletons, pagination, and new-session controls.
`SidebarSessionItem.tsx` for per-session desktop/mobile item rendering and session actions.
`SessionProviderIcon.tsx` for provider icon normalization.
`SidebarHeader.tsx`, `SidebarFooter.tsx`, `SidebarCollapsed.tsx`, and `SidebarModals.tsx` as dedicated typed UI surfaces.
This keeps rendering responsibilities local and significantly improves traceability.
- Convert shared UI primitives from JSX to TSX:
`src/components/ui/button.tsx`,
`src/components/ui/input.tsx`,
`src/components/ui/badge.tsx`,
`src/components/ui/scroll-area.tsx`.
These now provide typed props/variants (`forwardRef` where appropriate) while preserving existing class/behavior.
- Add shared app typings in `src/types/app.ts` for projects/sessions/websocket/loading contracts used by new hooks/components.
- Add global window declarations in `src/types/global.d.ts` for `__ROUTER_BASENAME__`, `refreshProjects`, and `openSettings`, removing implicit `any` usage for global integration points.
- Update `src/main.jsx` to import `App.tsx` and keep app bootstrap consistent with the TS migration.
- Update `src/components/QuickSettingsPanel.jsx` to self-resolve mobile state via `useDeviceSettings` (remove `isMobile` prop dependency), and update `src/components/ChatInterface.jsx` to render `QuickSettingsPanel` directly.
This reduces prop drilling and keeps quick settings colocated with chat UI concerns.
* refactor(sidebar): integrate settings modal into SidebarModals and update props
* fix(mobile): prevent menu tap from triggering unintended dashboard navigation
The mobile sidebar menu button redirects users to `cloudcli.ai/dashboard` when a
session was active. The redirect happened because
the menu was opened on `touchstart`, which mounted the sidebar before the touch
sequence completed; the follow-up tap/click then landed on the sidebar header
anchor.
This change rewrites mobile menu interaction handling in `MainContent.jsx` to
eliminate touch/click event leakage and ghost-click behavior.
Key changes:
- Added `suppressNextMenuClickRef` to guard against synthetic click events that
fire after a touch interaction.
- Added `openMobileMenu(event)` helper to centralize `preventDefault`,
`stopPropagation`, and `onMenuClick()` invocation.
- Added `handleMobileMenuTouchEnd(event)`:
- opens the menu on `touchend` instead of `touchstart`
- sets a short suppression window (350ms) for the next click.
- Added `handleMobileMenuClick(event)`:
- ignores/suppresses click events during the suppression window
- otherwise opens the menu normally.
- Updated all mobile menu button instances in `MainContent.jsx` (loading state,
no-project state, active header state) to use:
- `onTouchEnd={handleMobileMenuTouchEnd}`
- `onClick={handleMobileMenuClick}`
- Removed the previous `onTouchStart` path that caused premature DOM mutation.
Behavioral impact:
- Mobile sidebar still opens reliably with one tap.
- Tap no longer leaks to newly-mounted sidebar header links.
- Prevents accidental redirects while preserving existing menu UX.
* refactor(main-content): migrate MainContent to TypeScript and modularize UI/state boundaries
Replace the previous monolithic MainContent.jsx with a typed one and
extract focused subcomponents/hooks to improve readability, local state ownership,
and maintainability while keeping runtime behavior unchanged.
Key changes:
- Replace `src/components/MainContent.jsx` with `src/components/MainContent.tsx`.
- Add typed contracts for main-content domain in `src/components/main-content/types.ts`.
- Extract header composition into:
- `MainContentHeader.tsx`
- `MainContentTitle.tsx`
- `MainContentTabSwitcher.tsx`
- `MobileMenuButton.tsx`
- Extract loading/empty project views into `MainContentStateView.tsx`.
- Extract editor presentation into `EditorSidebar.tsx`.
- Move editor file-open + resize behavior into `useEditorSidebar.ts`.
- Move mobile menu touch/click suppression logic into `useMobileMenuHandlers.ts`.
- Extract TaskMaster-specific concerns into `TaskMasterPanel.tsx`:
- task detail modal state
- PRD editor modal state
- PRD list loading/refresh
- PRD save notification lifecycle
Behavior/compatibility notes:
- Preserve existing tab behavior, session passthrough props, and Chat/Git/File flows.
- Keep interop with existing JS components via boundary `as any` casts where needed.
- No intentional functional changes; this commit is structural/type-oriented refactor.
Validation:
- `npm run typecheck` passes.
- `npm run build` passes (existing unrelated CSS minify warnings remain).
* refactor(chat): split monolithic chat interface into typed modules and hooks
Replace the legacy monolithic ChatInterface.jsx implementation with a modular TypeScript architecture centered around a small orchestration component (ChatInterface.tsx).
Core architecture changes:
- Remove src/components/ChatInterface.jsx and add src/components/ChatInterface.tsx as a thin coordinator that wires provider state, session state, realtime WebSocket handlers, and composer behavior via dedicated hooks.
- Update src/components/MainContent.tsx to use typed ChatInterface directly (remove AnyChatInterface cast).
State ownership and hook extraction:
- Add src/hooks/chat/useChatProviderState.ts to centralize provider/model/permission-mode state, provider/session synchronization, cursor model bootstrap from backend config, and pending permission request scoping.
- Add src/hooks/chat/useChatSessionState.ts to own chat/session lifecycle state: session loading, cursor/claude/codex history loading, pagination, scroll restoration, visible-window slicing, token budget loading, persisted chat hydration, and processing-state restoration.
- Add src/hooks/chat/useChatRealtimeHandlers.ts to isolate WebSocket event processing for Claude/Cursor/Codex, including session filtering, streaming chunk buffering, session-created/pending-session transitions, permission request queueing/cancellation, completion/error handling, and session status updates.
- Add src/hooks/chat/useChatComposerState.ts to own composer-local state and interactions: input/draft persistence, textarea sizing and keyboard behavior, slash command execution, file mentions, image attachment/drop/paste workflow, submit/abort flows, permission decision responses, and transcript insertion.
UI modularization under src/components/chat:
- Add view/ChatMessagesPane.tsx for message list rendering, loading/empty states, pagination affordances, and thinking indicator.
- Add view/ChatComposer.tsx for composer shell layout and input area composition.
- Add view/ChatInputControls.tsx for mode toggles, token display, command launcher, clear-input, and scroll-to-bottom controls.
- Add view/PermissionRequestsBanner.tsx for explicit tool-permission review actions (allow once / allow & remember / deny).
- Add view/ProviderSelectionEmptyState.tsx for provider and model selection UX plus task starter integration.
- Add messages/MessageComponent.tsx and markdown/Markdown.tsx to isolate message rendering concerns, markdown/code rendering, and rich tool-output presentation.
- Add input/ImageAttachment.tsx for attachment previews/removal/progress/error overlay rendering.
Shared chat typing and utilities:
- Add src/components/chat/types.ts with shared types for providers, permission mode, message/tool payloads, pending permission requests, and ChatInterfaceProps.
- Add src/components/chat/utils/chatFormatting.ts for html decoding, code fence normalization, regex escaping, math-safe unescaping, and usage-limit text formatting.
- Add src/components/chat/utils/chatPermissions.ts for permission rule derivation, suggestion generation, and grant flow.
- Add src/components/chat/utils/chatStorage.ts for resilient localStorage access, quota handling, and normalized Claude settings retrieval.
- Add src/components/chat/utils/messageTransforms.ts for session message normalization (Claude/Codex/Cursor) and cached diff computation utilities.
Command/file input ergonomics:
- Add src/hooks/chat/useSlashCommands.ts for slash command fetching, usage-based ranking, fuzzy filtering, keyboard navigation, and command history persistence.
- Add src/hooks/chat/useFileMentions.tsx for project file flattening, @mention suggestions, mention highlighting, and keyboard/file insertion behavior.
TypeScript support additions:
- Add src/types/react-syntax-highlighter.d.ts module declarations to type-check markdown code highlighting imports.
Behavioral intent:
- Preserve existing chat behavior and provider flows while improving readability, separation of concerns, and future refactorability.
- Move state closer to the components/hooks that own it, reducing cross-cutting concerns in the top-level chat component.
* perf(project-loading): eliminate repeated Codex session rescans and duplicate cursor fetches
The staged changes remove the main source of project-load latency by avoiding repeated full scans of ~/.codex/sessions for every project and by removing redundant client-side cursor session refetches.
Server changes (server/projects.js):\n- Add a per-request Codex index reference in getProjects so Codex metadata is built once and reused across all projects, including manually added ones.\n- Introduce normalizeComparablePath() to canonicalize project paths (including Windows long-path prefixes and case-insensitive matching on Windows).\n- Introduce findCodexJsonlFiles() + buildCodexSessionsIndex() to perform a single recursive Codex scan and group sessions by normalized cwd.\n- Update getCodexSessions() to accept indexRef and read from the prebuilt index, with fallback index construction when no ref is provided.\n- Preserve existing session limiting behavior (limit=5 default, limit=0 returns all).
Client changes (src/hooks/useProjectsState.ts):\n- Remove loadCursorSessionsForProjects(), which previously triggered one extra /api/cursor/sessions request per project after /api/projects.\n- Use /api/projects response directly during initial load and refresh.\n- Expand projectsHaveChanges() to treat both cursorSessions and codexSessions as external session deltas.\n- Keep refresh comparison aligned with external session updates by using includeExternalSessions=true in sidebar refresh path.
Impact:\n- Reduces backend work from roughly O(projects * codex_session_files) to O(codex_session_files + projects) for Codex discovery during a project load cycle.\n- Removes an additional client-side O(projects) network fan-out for Cursor session fetches.\n- Improves perceived and actual sidebar project-loading time, especially in large session datasets.
* fix(chat): make Stop and Esc reliably abort active sessions
Problem
Stop requests were unreliable because aborting depended on currentSessionId being set, Esc had no actual global abort binding, stale pending session ids could be reused, and abort failures were surfaced as successful interruptions. Codex sessions also used a soft abort flag without wiring SDK cancellation.
Changes
- Add global Escape key handler in chat while a run is loading/cancellable to trigger the same abort path as the Stop button.
- Harden abort session target selection in composer by resolving from multiple active session id sources (current, pending view, pending storage, cursor storage, selected session) and ignoring temporary new-session-* ids.
- Clear stale pendingSessionId when launching a brand-new session to prevent aborting an old run.
- Update realtime abort handling to respect backend success=false responses: keep loading state when abort fails and emit an explicit failure message instead of pretending interruption succeeded.
- Improve websocket send reliability by checking socket.readyState === WebSocket.OPEN directly before send.
- Implement real Codex cancellation via AbortController + runStreamed(..., { signal }), propagate aborted status, and suppress expected abort-error noise.
Impact
This makes both UI Stop and Esc-to-stop materially more reliable across Claude/Cursor/Codex flows, especially during early-session windows before currentSessionId is finalized, and prevents false-positive interrupted states when backend cancellation fails.
Validation
- npm run -s typecheck
- npm run -s build
- node --check server/openai-codex.js
* refactor: tool components
* refactor: tool components
* fix: remove one-line logic from messagecomponent
* refactor(design): change the design of bash
* refactor(design): fix bash design and config
* refactor(design): change the design of tools and introduce todo list and task list.
* refactor(improvement):add memo on diffviewer, cleanup messsagecomponent
* refactor: update readme and remove unusedfiles.
* refactor(sidebar): remove duplicate loading message in SidebarProjectsState
* refactor(sidebar): move VersionUpgradeModal into SidebarModals
* refactor: replace individual provider logos with a unified SessionProviderLogo component
* fix(commands): restore /cost slash command and improve command execution errors
The /cost command was listed as built-in but had no handler, causing execution to
fall through to custom command logic and return 400 ("command path is required").
- Add a built-in /cost handler in server/routes/commands.js
- Return the expected payload shape for the chat UI (`action: "cost"`, token usage,
estimated cost, model)
- Normalize token usage inputs and compute usage percentage
- Add provider-based default pricing for cost estimation
- Fix model selection in command execution context so codex uses `codexModel`
instead of `claudeModel`
- Improve frontend command error handling by parsing backend error responses and
showing meaningful error messages instead of a generic failure
* fix(command-menu): correct slash command selection with frequent commands
When the “Frequently Used” section is visible, command clicks/hover could use a
UI-local index instead of the canonical `filteredCommands` index, causing the
wrong command to execute (e.g. clicking `/help` running `/clear`).
- map rendered menu items back to canonical command indices using a stable key
(`name + namespace/type + path`)
- use canonical index for hover/click selection callbacks
- deduplicate frequent commands from other grouped sections to avoid duplicate
rows and selection ambiguity
- keep and restore original inline comments, with clarifications where needed
* refactor(sidebar): update sessionMeta handling for session loading logic
- This fixes an issue where the sidebar was showing 6+ even when there were only 5 sessions, due to the hasMore logic not accounting for the case where there are exactly 6 sessions.
It was also showing "Show more sessions" even where there were no more sessions to load.
- This was because `hasMore` was sometimes `undefined` and the logic checked for hasMore !== false, which treated undefined as true.
Now we explicitly check for hasMore === true to determine if there are more sessions to load.
* refactor(project-watcher): add codex and cursor file watchers
* fix: chat session scroll to bottom error even when scrolled up
* fix(chat): clear stuck loading state across realtime lifecycle events
The chat UI could remain in a stale "Thinking/Processing" state when session IDs
did not line up exactly between view state (`currentSessionId`), selected route
session, pending session IDs, and provider lifecycle events. This was most visible
with Codex completion/abort flows, but the same mismatch risk existed in shared
handlers.
Unify lifecycle cleanup behavior in realtime handlers and make processing tracking
key off the active viewed session identity.
Changes:
- src/hooks/chat/useChatRealtimeHandlers.ts
- src/components/ChatInterface.tsx
- src/hooks/chat/useChatSessionState.ts
What changed:
- Added shared helpers in realtime handling:
- `collectSessionIds(...)` to normalize and dedupe candidate session IDs.
- `clearLoadingIndicators()` to consistently clear `isLoading`, abort UI, and status.
- `markSessionsAsCompleted(...)` to consistently notify inactive/not-processing state.
- Updated lifecycle branches to use shared cleanup logic:
- `cursor-result`
- `claude-complete`
- `codex-response` (`turn_complete` and `turn_failed`)
- `codex-complete`
- `session-aborted`
- Expanded completion/abort cleanup to include all relevant session IDs
(`latestMessage.sessionId`, `currentSessionId`, `selectedSession?.id`,
`pendingSessionId`, and Codex `actualSessionId` when present).
- Switched processing-session marking in `ChatInterface` to use
`selectedSession?.id || currentSessionId` instead of `currentSessionId` alone.
- Switched processing-session rehydration in `useChatSessionState` to use
the same active-view session identity fallback.
Result:
- Prevents stale loading indicators after completion/abort when IDs differ.
- Keeps processing session bookkeeping aligned with the currently viewed session.
- Reduces provider-specific drift by using one lifecycle cleanup pattern.
* fix(chat): stabilize long-history scroll-up pagination behavior
- fix top-pagination lockups by only locking when older messages are actually fetched
- make fetched older messages visible immediately by increasing `visibleMessageCount` on prepend
- prevent unintended auto-scroll-to-bottom during older-message loading and scroll restore
- replace state-based pagination offset with a ref to avoid stale offset/reload side effects
- ensure initial auto-scroll runs only after initial session load completes
- reset top-load lock/restore state and visible window when switching sessions
- loosen top-lock release near the top to avoid requiring a full down/up cycle
* refactor: Restructure files and folders to better mimic feature-based architecture
* refactor: reorganize chat view components and types
* feat(chat): move thinking modes, token usage pie, and related logic into chat folder
* refactor(tools): add agent category for Task tool
Add visual distinction for the Task tool (subagent invocation) by
introducing a new 'agent' category with purple border styling. This
separates subagent tasks from regular task management tools
(TaskCreate, TaskUpdate, etc.) for clearer user feedback.
Also refactor terminal command layout in OneLineDisplay to properly
nest flex containers, fixing copy button alignment issues.
* refactor(tools): improve Task tool display formatting
Update Task tool config to show cleaner subagent information in the UI.
Simplifies the input display by showing only the prompt when no
optional fields are present, reducing visual clutter. Updates title
format to "Subagent / {type}" for better categorization. Enhances
result display to better handle complex agent response structures with
array content types, extracting text blocks for cleaner presentation.
* fix: show auth url panel in shell only on mobile
- use static url: https://auth.openai.com/codex/device, for codex login.
- add an option for hiding the panel
* fix(chat): escape command name in regex to prevent unintended matches
* fix(chat): handle JSON parsing errors for saved chat messages
* refactor(chat): replace localStorage provider retrieval with prop usage in MessageComponent
* fix(chat): handle potential null content in message before splitting lines
* refactor(todo): update TodoListContentProps to include optional id and priority fields that are used in TodoList.jsx
* fix(watcher): ensure provider folders exist before creating watchers to maintain active watching
* refactor(chat): improve message handling by cloning state updates and improving structured message parsing
* refactor(chat): exclude currentSessionId from dependency array to prevent unnecessary reloading of messages
* refactor(useFileMentions): implement abort controller for fetch requests
* refactor(MessageComponent): add types
* refactor(calculateDiff): optimize LCS algorithm for improved diff calculation
* refactor(createCachedDiffCalculator): use both newStr and oldStr as cache keys
* refactor(useSidebarController): manage project session overrides in local state
* refactor(ScrollArea): adjust ref placement and className order
* fix: type annotations
* refactor(ChatInputControls): update import statement for ThinkingModeSelector
* refactor(dateUtils): update type annotation for formatTimeAgo function
* refactor(ToolRenderer): ensure stable hook order
* refactor(useProjectsState): normalize refreshed session metadata to maintain provider stability; use getProjectSessions helper for session retrieval.
* refactor(useChatComposerState): improve input handling and command execution flow
* refactor(useChatRealtimeHandlers): normalize interactive prompt content to string for consistent ChatMessage shape
* refactor(OneLineDisplay): improve clipboard functionality with fallback for unsupported environments
* refactor(QuickSettingsPanel): simplify state management by removing localIsOpen and using isOpen directly
* refactor(ChatMessagesPane): use stable message key
* refactor:: move AssistantThinkingIndicator component to its own file
* refactor(ChatMessagesPane): extract message key generation logic to a utility function
* refactor(SidebarModals): move normalizeProjectForSettings into utils file
* refactor(ToolConfigs): use optional chaining for content retrieval
* fix(chat): stabilize provider/message handling and complete chat i18n coverage
Unify provider typing, harden realtime message effects, normalize tool input
serialization, and finish i18n/a11y updates across chat UI components.
- tighten provider contracts from `Provider | string` to `SessionProvider` in:
- `useChatProviderState`
- `useChatComposerState`
- `useChatRealtimeHandlers`
- `ChatMessagesPane`
- `ProviderSelectionEmptyState`
- refactor `AssistantThinkingIndicator` to accept `selectedProvider` via props
instead of reading provider from local storage during render
- fix stale-closure risk in `useChatRealtimeHandlers` by:
- adding missing effect dependencies
- introducing `lastProcessedMessageRef` to prevent duplicate processing when
dependencies change without a new message object
- standardize `toolInput` shape in `messageTransforms`:
- add `normalizeToolInput(...)`
- ensure all conversion paths produce consistent string output
- remove mixed `null`/raw/stringified variants across cursor/session branches
- harden tool display fallback in `CollapsibleDisplay`:
- default border class now falls back safely for unknown categories
- improve chat i18n consistency:
- localize hardcoded strings in `MessageComponent`
(`permissions.*`, `interactive.*`, `thinking.emoji`, `json.response`,
`messageTypes.error`)
- localize button titles in `ChatInputControls`
(`input.clearInput`, `input.scrollToBottom`)
- localize provider-specific empty-project prompt in `ChatInterface`
(`projectSelection.startChatWithProvider`)
- localize repeated “Start the next task” prompt in
`ProviderSelectionEmptyState` (`tasks.nextTaskPrompt`)
- add missing translation keys in all supported chat locales:
- `src/i18n/locales/en/chat.json`
- `src/i18n/locales/ko/chat.json`
- `src/i18n/locales/zh-CN/chat.json`
- new keys:
- `input.clearInput`
- `input.scrollToBottom`
- `projectSelection.startChatWithProvider`
- `tasks.nextTaskPrompt`
- improve attachment remove-button accessibility in `ImageAttachment`:
- add `type="button"` and `aria-label`
- make control visible on touch/small screens and focusable states
- preserve hover behavior on larger screens
Validation:
- `npm run typecheck`
* fix(chat): sync quick settings state and stabilize thinking toggle
Synchronize useUiPreferences instances via custom sync events and storage listeners so Quick Settings updates apply across UI consumers immediately.
Also hide standalone thinking messages when showThinking is disabled, while preserving hook order to avoid Rendered fewer hooks runtime errors.
* refactor(validateGitRepository): improve directory validation for git work tree
* refactor(GitPanel): clear stale state on project change and improve error handling
* refactor(git): use spawnAsync for command execution and improve commit log retrieval
* fix: iOS pwa bottom margin
* fix: pass diff information to code editor
* refactor(sidebar): remove touch event handlers from project and session items
* bumping node to v22
* Release 1.17.0
---------
Co-authored-by: Haileyesus <something@gmail.com>
Co-authored-by: simosmik <simosmik@gmail.com>
Update .env.example with comprehensive CLI command documentation and
clearer DATABASE_PATH configuration comments. Enhance README.md with
restructured installation guide featuring new cloudcli commands,
detailed PM2 background service setup instructions, and improved
organization of global installation benefits and restart procedures.
Add CLI command reference showing cloudcli start, status, help, and
version commands. Expand PM2 section with separate subsections for
installation, service startup, and auto-start configuration.
