## Summary
- add server-backed speech capabilities and transcription endpoints plus
UI settings for speech configuration
- add push-to-talk prompt voice input with microphone controls,
transcription insertion, and browser capability gating
- keep prompt controls aligned by restoring right-side nav placement and
moving the mic beside the expand control
# feat(i18n): Hebrew locale + full RTL support
## Summary
This PR adds full Hebrew (he) locale support to the UI, including a
complete translation of all user-facing strings and comprehensive RTL
layout support across all components.
## What was done
### Hebrew translation
- Full translation of all i18n message files for the `he` locale (17
translation files)
- Registered the language in the i18n system and the language picker
### RTL support
- Automatic direction detection (`dir="rtl"`) when Hebrew is selected
- Replaced physical CSS properties (`left`/`right`) with logical
equivalents (`inline-start`/`inline-end`) across the project
- Fixed resize direction, file path alignment, and textarea padding
- Fixed navigation button positioning in textarea for RTL
- Fixed scrollbar direction in RTL
- Fixed code block direction and selector alignment
- Fixed Monaco editor direction in the file viewer
- Auto-detect text direction in reasoning block (`dir="auto"` +
`unicode-bidi: plaintext`)
### Adapted components
- `session-layout` — sidebar and resize handle
- `prompt-input` — text direction and buttons
- `message-base` — message blocks and reasoning
- `message-timeline` — timeline bar
- `right-panel` — right side panel
- `tool-call` — tool call display
- `settings-screen` — settings page
- `selector` — selection component
- `instance-shell` — main shell
## New files
```
packages/ui/src/lib/i18n/messages/he/
advancedSettings.ts
app.ts
commands.ts
dialogs.ts
filesystem.ts
folderSelection.ts
index.ts
instance.ts
loadingScreen.ts
logs.ts
markdown.ts
messaging.ts
remoteAccess.ts
session.ts
settings.ts
time.ts
toolCall.ts
```
## Suggested testing
- Switch language to Hebrew and verify all strings are translated
- Verify RTL layout is correct across all screens (session, settings,
file viewer)
- Verify that English text inside a reasoning block is displayed LTR
- Switch back to English and verify everything returns to LTR
---------
Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com>
Co-authored-by: Shantur Rathore <i@shantur.com>
## Summary
- defer locale and overlay loading work away from the first critical
render path
- seed locale state from the bootstrap preload so the first render can
use the preloaded language immediately
- keep bootstrap cache and locale fallback behavior consistent on
subsequent launches
## Testing
- npm run build --workspace @codenomad/ui
## Summary
- split the right panel, picker, and tool call secondary viewers into
smaller deferred chunks
- release hidden right-panel file buffers and stop tracking static
tool-call scrollers when they are not needed
- keep this branch focused on the remaining secondary viewer chunking
work now that the Monaco-specific chunking moved into PR 215
## Testing
- npm run build --workspace @codenomad/ui
## What and why
CodeNomad had no RTL (right-to-left) support, so users writing in Hebrew
or Arabic would see their messages displayed left-to-right — misaligned
text, broken reading flow, wrong punctuation placement.
This PR adds automatic direction detection to all elements that display
user or model text. The browser detects direction from the first strong
character in each text block: Hebrew/Arabic → RTL, Latin/code → LTR. No
configuration needed — it just works per message, per paragraph.
## Technical notes
The natural fix is `dir="auto"` on the containing elements. However,
Chromium does not propagate direction detection from a parent `<div>`
into its `<p>` children — so Hebrew inside `<p>` rendered via
`innerHTML` (as markdown is) was still detected as LTR. The fix is to
apply `unicode-bidi: plaintext` via CSS directly on the block-level
elements (`p`, `li`, headings, etc.), which has the same auto-detection
semantics but applies per element.
## Summary
- Add `dir="auto"` to all elements containing user-generated or
model-generated text (message content, prompt input, session names, tool
outputs) so the browser auto-detects text direction
- Add `unicode-bidi: plaintext` via CSS to markdown block elements (`p`,
`li`, headings, `blockquote`, `td`/`th`) to fix per-paragraph RTL
detection in Chromium (where `dir="auto"` on a parent div does not
recurse into block children)
- Convert physical CSS properties to logical equivalents in
`markdown.css`: `border-left` → `border-inline-start`, `padding-left` →
`padding-inline-start`, `text-align: left` → `text-align: start`,
`margin-left` → `margin-inline-start`
## Affected components
- `markdown.tsx` — main markdown renderer
- `message-part.tsx` — text part wrapper and plain-text fallback
- `message-item.tsx` — message body and error blocks
- `prompt-input.tsx` — user input textarea
- `session-list.tsx` — session titles in sidebar
- `session-rename-dialog.tsx` — session rename input
- `instance-welcome-view.tsx` — Resume Session dialog
- `tool-call/markdown-render.tsx` — tool output markdown fallback
- `tool-call/ansi-render.tsx` — ANSI output
- `tool-call/diagnostics-section.tsx` — diagnostic messages
## Test plan
- [ ] Send a Hebrew-only message → text right-aligned
- [ ] Send a mixed Hebrew + English message → correct per-paragraph
direction
- [ ] Message containing a code block → code stays LTR
- [ ] Type Hebrew in the prompt textarea → input flows right-to-left
- [ ] Hebrew session name in sidebar → right-aligned
- [ ] Hebrew session name in Resume Session dialog → right-aligned
🤖 Generated with [Claude Code](https://claude.com/claude-code)
---------
Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com>
## Summary
- lazy-load the markdown and diff render paths so they stop inflating
initial UI startup work
- move shared text rendering helpers out of the markdown path and keep
diff rendering on the deferred path
- defer the Monaco secondary viewers so the markdown and diff path no
longer keeps that work in the main bundle
## Follow-ups
- related fork follow-up: Pagecran/CodeNomad#1
- that follow-up is now independent on dev and only keeps the remaining
right panel, picker, and tool-call secondary chunking work
## Testing
- npm run typecheck --workspace @codenomad/ui
- npm run build --workspace @codenomad/ui
## What
Fix slash command execution so `[pasted #N]` placeholders are resolved
before calling `session.command`, matching normal prompt send behavior.
## Why
When pasting long text into a slash command (e.g. `/some-command [pasted
#1]`), the UI previously bypassed `resolvePastedPlaceholders(...)` for
known slash commands and sent the literal placeholder text as command
arguments.
## Changes
- Resolve pasted placeholders (and other prompt placeholders handled by
`resolvePastedPlaceholders`) in slash-command arguments before
`executeCustomCommand(...)`.
- Remove *consumed* pasted-text attachments (those referenced by
placeholders in the slash-command args) so they don’t linger for the
next prompt.
Fixes#234.
## Notes
- I attempted `npm run typecheck --workspace @codenomad/ui` locally but
the workspace dependencies aren’t installed in this bot environment, so
it fails with missing-module errors. CI should validate with a full
install.
--
Yours,
[CodeNomadBot](https://github.com/NeuralNomadsAI/CodeNomad)
Co-authored-by: Shantur Rathore <i@shantur.com>
## Summary
- restore the GitHub and Discord links on the folder picker in the Tauri
app
- open those links through the desktop opener bridge instead of relying
on browser-only navigation behavior
- include the capability/schema updates needed for the opener path
## Testing
- npm run typecheck --workspace @codenomad/ui
- cargo check --manifest-path packages/tauri-app/src-tauri/Cargo.toml
This PR prevents OS notification spam from spawned subagent sessions by
skipping OS-level notifications for any session that is a child thread
(`parentId !== null`).
What changed
- `packages/ui/src/stores/session-events.ts`
- Added `isChildSession(...)` +
`shouldSendOsNotificationForSession(...)`
- Applied the check to OS notifications emitted from:
- `handleSessionIdle(...)`
- `handlePermissionUpdated(...)`
- `handleQuestionAsked(...)`
- If a session is not yet hydrated in the client store, we
conservatively *do not* emit an OS notification (avoids early subagent
spam).
Why
- Subagent sessions are represented as child sessions in the UI thread
model; OS notifications were previously emitted for all sessions
indiscriminately.
Testing
- Not run here: `bun run typecheck` fails in this environment due to
missing installed deps/types (e.g. `solid-js`).
Closes#228
--
Yours,
[CodeNomadBot](https://github.com/NeuralNomadsAI/CodeNomad)
Co-authored-by: Shantur Rathore <i@shantur.com>
### Summary of Improvements
This PR replaces the custom `IntersectionObserver`-based virtualization
with the `virtua` library to significantly improve rendering performance
and UI responsiveness.
### 🚀 Performance Results
Verified using `session-performance.test.ts`:
- **Rendering**: 2000 messages rendered in **16.90ms**.
- **Huge Conversation**: 10,000 messages processed in **0.80ms**.
- **Session Switching**: Average switch time reduced to **0.58ms**
(virtually zero lag).
### 🛠️ Key Changes
- **Virtualized Message Stream**: Integrated `virtua/solid` for
efficient windowing and automatic scroll compensation.
- **Floating Scroll Controls**: Applied `position: absolute` and
`pointer-events: none` to the list controls to ensure
scroll-to-top/bottom buttons float correctly over the message area
without blocking interactions.
- **Package Synchronization**: Updated `virtua` and SDK dependencies,
with a fully synchronized `package-lock.json` for stable builds.
### 🎥 UI Verification
https://github.com/user-attachments/assets/24e483a3-8be6-4ac4-a431-d719f2015f4e
- **Smooth Scrolling**: Verified that rendering gaps are eliminated
during fast scrolls.
- **Position Retention**: Scroll positions are preserved when switching
between sessions.
> [!NOTE]
> Detailed performance gains and layout fixes are isolated to the
`virtua` implementation and core package updates, following the
requested cleanup.
---------
Co-authored-by: Shantur Rathore <i@shantur.com>
Reconcile optimistic user messages by replacing the oldest synthetic pending message when the server-backed message arrives. Stop sending prompt part ids and rely on message-level replacement so v1.2.25 validation passes without duplicating optimistic content.
Disable follow-mode virtualization churn and simplify reasoning header layout so streaming thinking blocks stop nudging the scroll position while the list is pinned to bottom.
Keep visible rows mounted during follow-up measurements and clear stale refs so async message rendering no longer flickers or measures detached blocks. Coalesce per-item render notifications so content-heavy rows only trigger one remeasurement per frame.
Complete re-review of PR #188 (commits 224cab6 feature + 2c27fc5 perf/i18n follow-up). Gatekeeper focus: standards, correctness, perf/complexity, and translation completeness.
What this changes (pre -> post)
Pre: timeline primarily navigation/hover preview; bulk delete selection message-level and token metrics tied to backend assistant output tokens (missing tool payload weight).
Post: segment-level timeline selection + range (Shift) + toggle (Ctrl/Meta) + mobile long-press; histogram ribs overlay showing relative + absolute (~10k cap) token weight; assistant-turn grouping to avoid adjacency bugs; bulk-delete toolbar shows Before / Selection / After token pills.
Code standards / correctness
OK: Solid signal/memo/effect patterns with cleanup; no obvious lifecycle leaks. Grouping avoids adjacency overlap by mapping messageId to turns.
Fix: selection-id stability is mitigated by pruning stale ids after segment rebuilds; long term stable ids from part ids/toolPartIds remain recommended.
Fix: token counts now share getPartCharCount in both x-ray overlay and bulk-delete toolbar, keeping estimates consistent with live store updates.
Performance / complexity
OK: O(n^2) hotspots removed for liveSegmentChars and selectedTokenTotal. groupRole + deleteUpTo hover checks now memoize messageId sets/maps.
Note: getPartCharCount can be heavy for large tool payloads but remains gated behind selection mode.
CSS / UI integration
Fix: x-ray token label now uses theme tokens instead of hard-coded colors. Delete toolbar now uses menu-based controls with selection-mode toggle.
i18n
Fix: selection hint now renders Cmd/Ctrl via localized modifier placeholder; all locales updated.
