- Add local Whisper STT via Transformers.js with Web Worker (no UI freeze)
- Default sttProvider to 'local' in Electron (browser STT unavailable)
- Fix infinite toast loop: stop auto-restart on network errors
- Add retry limit with exponential backoff for transient STT errors
- Append voice transcript to input field (append-inline), not replace
- Add model catalog with download/load button in Voice Settings
When notificationMode is 'hidden-only', the isWindowFocused check now
happens before summarization and template resolution, not after. This
prevents costly Zen API calls for notifications that would be skipped
anyway, and eliminates stale notifications arriving after the user has
already read the response and switched away.
C.UTF-8 does not exist on macOS, causing the terminal to fall back
to ASCII encoding when LANG is unset. Use en_US.UTF-8 on macOS and
C.UTF-8 on Linux. Also set LC_CTYPE with the platform-appropriate
value (UTF-8 on macOS, C.UTF-8 on Linux) for additional safety.
Subagent completion notifications respect notification settings
Permission auto-accept now suppresses notifications for known child sessions
Server notification checks use OpenCode session parent semantics
Add a Stashes dialog with create, apply, pop, and drop actions
Include untracked files automatically when stashing
Show file counts for current changes and stash entries
Combine fetch, pull with rebase, and push into one sync action
Keep remote dropdown focused on safe fetch actions
Block sync when uncommitted changes would conflict with rebase
* fix: add concurrency controls for multiple sessions using the same provider
Adds OS-inspired scheduling primitives (from HiveMind/AIMD research) to prevent
concurrent sessions from the same provider from experiencing slowdowns, random
stops, and cascading failures.
Server-side:
- Health check skips OpenCode restart when sessions are actively busy — a busy
server under concurrent load can fail the health check timeout without being
dead. Staleness guard forces restart if unhealthy+busy persists >2 minutes.
- Upstream SSE stall timeout scaled from 20s to 60s to avoid unnecessary
reconnections when multiple sessions are waiting for LLM responses.
Client-side (HiveMind primitives, arXiv:2604.17111):
- Transparent retry with exponential backoff (1s→2s→4s, max 32s) for
429/502/503/504 errors — the #1 most effective primitive from the paper.
- Circuit breaker: opens after 3 consecutive retryable errors, cooldown
doubles each trip (30s→60s→120s, capped 128s), matching TCP AIMD.
- Per-provider session tracking with TTL eviction (1h idle sweep).
- Fetch-level retry gated on AbortError/TypeError only (not DNS failures).
Refs github-code-review skill findings (all 8 issues resolved).
* fix: use definite assignment assertion for response variable
Fixes TS2454: Variable 'response' is used before being assigned
in strict mode. The for-loop body always assigns it on every path
that reaches the post-loop code, but TS can't prove that.
* fix: add cleanupSession to error paths and remove unreachable code
P1 fixes (Greptile review):
- cleanupSession called on fetch error throw path
- cleanupSession called on non-retryable HTTP error throw path
- Removed unreachable post-loop code (loop always terminates via return or throw)
Adds explicit post-loop throw to satisfy TypeScript strict return check.
* fix: address Greptile review feedback on concurrent session controls
Removes client-side session tracking that leaked on normal completion paths.
The session tracking was redundant — the server-side health check already reads from
sessionRuntime.getSessionActivitySnapshot() for busy-session detection.
Changes:
- Remove activeSessions Set and all session-tracking functions from provider-tracker
- Remove trackSessionStarted/cleanupSession calls from client.ts
- Remove unreachable (response as Response) block after retry loop
- Make upstreamStallTimeoutMs conditional: 60s when >1 sessions, 20s otherwise
Refs #1069
* fix: enforce dynamic concurrency safeguards
---------
Co-authored-by: Bohdan Triapitsyn <artmore@protonmail.com>
* fix(server): increase WS buffer/replay limits and add backpressure warning
During long-running agent sessions (e.g. ultrawork loops with many tool
calls), the browser WebSocket client can briefly fall behind the server.
When the outbound buffer exceeds the limit, the server force-disconnects
with close code 1013, and the small replay buffer (512 events) is
insufficient to recover all missed events — leaving the UI permanently
stalled.
Changes:
- Raise MESSAGE_STREAM_WS_MAX_BUFFERED_BYTES from 4 MB to 16 MB to
tolerate larger bursts without disconnecting
- Add MESSAGE_STREAM_WS_BACKPRESSURE_WARN_BYTES (12 MB) threshold that
sends a one-shot "backpressure" frame to the client before the hard
disconnect, giving it a chance to shed low-priority updates
- Raise MESSAGE_STREAM_GLOBAL_REPLAY_LIMIT from 512 to 2048 so more
events survive brief reconnection gaps
- Add tests for the backpressure warning behavior (emit, dedup, reset)
* fix(ui): batch event flushes under backpressure
---------
Co-authored-by: Bohdan Triapitsyn <artmore@protonmail.com>
* fix: cross-verify update API claims against npm registry
* Update packages/web/server/lib/package-manager.js
Signed-off-by: Islam Nofl <islamnofl.official@gmail.com>
* fix: show live server version in AboutDialog instead of stale build-time constant
* fix: add comment to empty catch block to satisfy lint no-empty rule
* fix: preserve live about dialog version in electron
* fix: scope update checks by runtime
---------
Signed-off-by: Islam Nofl <islamnofl.official@gmail.com>
Co-authored-by: Bohdan Triapitsyn <artmore@protonmail.com>
