Releases: off-grid-ai/OGAD
Releases · off-grid-ai/OGAD
Release list
Off Grid AI Desktop 0.0.42
Highlights
Maintenance release. No user-facing changes this build.
What's Changed
- chore: bump version to 0.0.42 for the stable release (8d18fe2)
- test(e2e): refresh screenshots from the production-build run (4a885bc)
- test(e2e): make the last two specs honest on the PRODUCTION target (57886fc)
- test(e2e): seed a cached Pro license so pro specs can verify the PRODUCTION build (ad12a6c)
- test(e2e): one launch seam so the suite can run against the PRODUCTION build (8217bb2)
Full Changelog: v0.0.40...v0.0.42
Off Grid AI Desktop 0.0.41-beta.77
Highlights
New
- Llama-server falls back to a free port when :8439 is taken (increment 2)
- Pure free-port chooser with gradual +1 fallback (foundation)
Fixes
- Follow-ups from CodeRabbit review of the merged port-fallback work
- Gateway self-report routes advertise the LIVE bound port
- Handle startModelServer rejection on the promise, not a try/catch
- Fall back for ANY port holder, not only a recognized llama owner
- Auto-fallback for the gateway (7878) and media (7879) ports
What's Changed
- test(e2e): make clipboard quick-open focus the app before its native hotkey (e4ab953)
- chore: satisfy explicit-return-type and prettier on the new e2e helpers (6b5033d)
- ci(e2e): retry in CI, and correct the advisory-gate record (ab9bbff)
- test(e2e): skip fixed-port specs when another process owns the engine ports (af1c382)
- test(e2e): cover Settings sections + resource mode (the surface that had none) (f930b35)
- test(e2e): repair meeting-transcription nav + consolidate onboarding into one helper (b43b19a)
- test(e2e): repair settings-residency — reach Model memory and derive its selectors (5b2f6ce)
- test(e2e): fix Settings tour specs — model the single-open accordion (585f388)
- refactor(settings): extract RESIDENCY_ROWS into a pure single-source-of-truth module (cd871cd)
- chore(hooks): full product name 'Off Grid AI Desktop' in the pre-push e2e note (CodeRabbit #68) (74306b4)
- ci+hooks: e2e is advisory (not blocking) until specs self-skip + headless Electron is stable (7ee7286)
- ci: run e2e as its own parallel job (xvfb), not appended to verify (e6df3be)
- ci+hooks: gate on the full Playwright e2e (df38bd9)
- fix(ports): follow-ups from CodeRabbit review of the merged port-fallback work (fd69cd7)
- chore(bench): bound each request + fail on absent prompt-token usage (1ab50a4)
- fix(ports): gateway self-report routes advertise the LIVE bound port (0734319)
- fix(ports): handle startModelServer rejection on the promise, not a try/catch (e27ef5d)
- fix(ports): fall back for ANY port holder, not only a recognized llama owner (d1cdb10)
- fix(ports): auto-fallback for the gateway (7878) and media (7879) ports (4891bf5)
- feat(net): llama-server falls back to a free port when :8439 is taken (increment 2) (3ef22dc)
- feat(net): pure free-port chooser with gradual +1 fallback (foundation) (244a490)
- chore(bench): chat prompt-eval benchmark (bench-chat.mjs) (e80dc59)
- test(e2e): refresh Google client-setup screenshot (Task 2 evidence) (e635f8b)
- test(e2e): screenshot the Replay enable/disable capture control (Task 4 evidence) (3363c12)
- test(e2e): screenshot the capture opt-in control in Settings (Task 1 evidence) (23193c4)
Full Changelog: v0.0.41-beta.76...v0.0.41-beta.77
Off Grid AI Desktop 0.0.40
Highlights
Maintenance release. No user-facing changes this build.
What's Changed
- No changes since the previous release.
Full Changelog: v0.0.41-beta.77...v0.0.40
Off Grid AI Desktop 0.0.41-beta.76
Highlights
New
- Clean, awaitable unload - free the model port without a force-quit
- Use the model's trained context window as the ceiling, not a fixed cap
- Inline STT model picker (view + change) in meetings
- Expose active STT engine/model provenance (core)
Fixes
- Explicit null check for the in-flight init in unload (SonarCloud reliability)
- Wrap composer chips instead of overflowing the input box
- Idle (inter-token) stream timeout, not total-duration - long answers survive
- Don't yank the view down while streaming when the user scrolls up
- Stop capping the agentic answer at 1024 tokens - inherit the max-output setting
- Default max output to auto (until EOS / window fills), not a 2048 cap
- Reattach restores the image-gen progress panel on remount
What's Changed
- fix(llm): explicit null check for the in-flight init in unload (SonarCloud reliability) (0eacf94)
- docs(gaps): log 4 deferred edges from desktop-pro #32 review (Gitar) (a0a3972)
- docs(gaps): log 2 deferred items from PR #60 review (bench resilience, provenance fallback) (4134368)
- review: address CodeRabbit findings on the release PR (805280c)
- chore(test): set the coverage floor to a uniform 85% across all metrics (486a84f)
- fix(chat): wrap composer chips instead of overflowing the input box (3123a8b)
- test(settings): label the ctx/max-output selects + update the response-limit test (932ff51)
- fix(llm): idle (inter-token) stream timeout, not total-duration — long answers survive (e7aef2f)
- fix(chat): don't yank the view down while streaming when the user scrolls up (3f163f6)
- fix(tools): stop capping the agentic answer at 1024 tokens — inherit the max-output setting (f9ead97)
- fix(llm): default max output to auto (until EOS / window fills), not a 2048 cap (63cc87d)
- feat(engine): clean, awaitable unload — free the model port without a force-quit (e2ce09c)
- feat(llm): use the model's trained context window as the ceiling, not a fixed cap (ab9af57)
- fix(chat): reattach restores the image-gen progress panel on remount (78b2b96)
- chore(bench): vision-only capture benchmark with batch sweep + quality judge (3889417)
- feat(transcription): inline STT model picker (view + change) in meetings (ef05f07)
- feat(transcription): expose active STT engine/model provenance (core) (a5918a5)
Full Changelog: v0.0.41-beta.75...v0.0.41-beta.76
Off Grid AI Desktop 0.0.41-beta.75
Highlights
Maintenance release. No user-facing changes this build.
What's Changed
- ci(coverage): calibrate Linux platform floors (d80e792)
Full Changelog: v0.0.41-beta.74...v0.0.41-beta.75
Off Grid AI Desktop 0.0.41-beta.74
Highlights
New
- Add semantic toast feedback
- Unify capture and model processing
- Surface actionable vision readiness
Fixes
- Skip markdown definitions during speech
- Recover malformed and transient state
- Preserve runtime data shapes
- Use canonical app support path
- Tolerate unavailable model summary IPC
- Share memory IPC records
- Roll back failed pipeline updates
- Stabilize vision setup recovery
- Preserve successful terminal state
- Refresh vision readiness from events
- Tolerate pro startup readiness gaps
- Keep jobs alive across navigation
- Preserve producer relationship projections
- Resolve packaged renderer from app root
- Prevent duplicate first-open microphone capture
What's Changed
- chore(types): remove obsolete baseUrl aliases (ab4fe2d)
- chore(types): check cross-layer renderer integrations (b1b8c53)
- fix(voice): skip markdown definitions during speech (25c2b9c)
- fix(renderer): recover malformed and transient state (6afa906)
- fix(contracts): preserve runtime data shapes (a395674)
- fix(scripts): use canonical app support path (e098219)
- docs(e2e): refresh final Pro evidence (9b76cb1)
- test(pro): accept merged capture settings (34cb663)
- test(e2e): isolate real engine model settings (b376c34)
- test(pro): make native acceptance time-safe (bd89de8)
- test(core): close admission and renderer review gaps (1e11415)
- fix(chat): tolerate unavailable model summary IPC (3a1b9a2)
- fix(types): share memory IPC records (52a81fb)
- chore(lint): type database records (00d3f76)
- chore(lint): type service integration boundaries (1691902)
- chore(lint): type model request boundaries (bf306eb)
- chore(lint): type licensing responses (aafbf9d)
- chore(lint): type shared product boundaries (57c15a1)
- chore(lint): type command and overlay primitives (ee4d711)
- chore(lint): type renderer menu primitives (342642d)
- refactor(release): isolate checklist generator (ce0e39c)
- chore(lint): type renderer service boundaries (16ca006)
- refactor(build): isolate signing operations (f0a6ed5)
- refactor(smoke): centralize acceptance harness (cd2aec3)
- test(integration): cover renderer and entity boundaries (81eb34f)
- test(e2e): clean up native voice acceptance (98cc7bd)
- chore(lint): type renderer visual primitives (71474db)
- refactor(scripts): isolate vision verification commands (7778111)
- refactor(scripts): isolate model download commands (ec6e25b)
- refactor(scripts): define command boundaries (32c3a50)
- refactor(tts): define worker boundary (7966c84)
- fix(settings): roll back failed pipeline updates (118c5e8)
- fix(capture): stabilize vision setup recovery (1cf1954)
- fix(imagegen): preserve successful terminal state (fcd98ce)
- test(ui): classify toast journey as integration (6944c65)
- feat(ui): add semantic toast feedback (26ea753)
- test(navigation): distinguish capture readiness listener (35f8717)
- test(imagegen): exercise durable job owner (9abe77b)
- fix(capture): refresh vision readiness from events (9517b3f)
- test(models): preserve live fixed-port owner (3b5a10d)
- test(settings): follow the processing control label (7164013)
- test(setup): supply the capture vision boundary (f5dab55)
- fix(capture): tolerate pro startup readiness gaps (b2653aa)
- test(pro): exercise native utility windows (14c7c4d)
- feat(settings): unify capture and model processing (f5b29db)
- fix(imagegen): keep jobs alive across navigation (c05665d)
- fix(entities): preserve producer relationship projections (015267e)
- feat(capture): surface actionable vision readiness (ee1baf6)
- fix(windows): resolve packaged renderer from app root (671c5e9)
- fix(voice): prevent duplicate first-open microphone capture (ea79197)
Full Changelog: v0.0.41-beta.73...v0.0.41-beta.74
Off Grid AI Desktop 0.0.41-beta.73
Highlights
New
- Allow confirmed rollback to signed releases
- Require consent for manual downloads
Fixes
- Beta channel must be 'beta' for prerelease discovery (not 'latest')
What's Changed
- test(update): synchronize packaged evidence with UI state (acc4eb4)
- test(update): cover signed release policy through service (3e41037)
- test(update): replace mockist checks with user journeys (3e43d52)
- test(update): exercise real rollback download start (ce113dc)
- test(update): verify rollback in packaged app (3e75cb9)
- docs(testing): require user-behavior integration tests (679084f)
- feat(update): allow confirmed rollback to signed releases (9a2e46a)
- feat(update): require consent for manual downloads (cf3ed31)
- fix(update): beta channel must be 'beta' for prerelease discovery (not 'latest') (6d64b4d)
Full Changelog: v0.0.41-beta.72...v0.0.41-beta.73
Off Grid AI Desktop 0.0.41-beta.72
Highlights
Fixes
- Normalize asar entry separators so build-win passes the inventory gate
What's Changed
- test(packaging): cover nested .app bundle with Windows separators (review follow-up) (f7449f9)
- fix(packaging): normalize asar entry separators so build-win passes the inventory gate (b194186)
Full Changelog: v0.0.41-beta.71...v0.0.41-beta.72
Off Grid AI Desktop 0.0.41-beta.71
Highlights
Fixes
- Stop Settings sections zooming/stretching on open (framer layout scale)
What's Changed
- test(e2e): Settings section drill-in/out + visible motion evidence (997a4ac)
- fix(ui): stop Settings sections zooming/stretching on open (framer layout scale) (58d1348)
Full Changelog: v0.0.41-beta.70...v0.0.41-beta.71
Off Grid AI Desktop 0.0.41-beta.70
Highlights
Fixes
- Stop offering an older stable as an update to a beta build
What's Changed
- fix(update): stop offering an older stable as an update to a beta build (97ae501)
Full Changelog: v0.0.41-beta.69...v0.0.41-beta.70