Skip to content

0.0.39-beta.66

Pre-release
Pre-release

Choose a tag to compare

View all tags
@github-actions github-actions released this 16 Jul 12:48
v0.0.39-beta.66
ea30213
This commit was created on GitHub.com and signed with GitHub’s verified signature.
GPG key ID: B5690EEEBB952194
Verified
Learn about vigilant mode.

Highlights

New

  • Tell non-Mac users Pro is coming soon on the upgrade screen
  • Device-aware copy + gate Pro to "coming soon" off macOS
  • Dev control to stop/restart the local model server
  • Cross-platform binary resolution + Windows build pipeline

Fixes

  • Set the Windows app icon (was defaulting to the Electron icon)
  • Pin transformers cacheDir to a writable dir (was re-downloading every run)
  • Fetch a current sd asset (cpu x64), image binary was silently missing
  • Budget tool schemas to context + surface server errors (B2/B3)
  • Pin bundled llama-server to b9838 to match the macOS engine
  • Fetch-win-binaries crashed on cosmetic listing, failing after copy
  • Guarantee model is loaded before ready + cross-platform orphan cleanup

What's Changed

  • chore(pr): address review — CI credential hygiene, docs, brand copy (c1b7ad8)
  • test(tools): add effectiveContextSize to the llm mock (fixes CI test gate) (4ef027f)
  • docs(win): README lists Windows as a supported platform (81509ff)
  • ci(win): publish a permanent Windows download link (parity with Mac) (1b04db3)
  • ci(win): fold a Windows build into release.yml (versioned, published, auto-updating) (78c577d)
  • feat(pro): tell non-Mac users Pro is coming soon on the upgrade screen (3d79573)
  • refactor(models): remove the dev-only stop/restart model-server control (bccc194)
  • fix(win): set the Windows app icon (was defaulting to the Electron icon) (e59c30c)
  • fix(embeddings): pin transformers cacheDir to a writable dir (was re-downloading every run) (947d914)
  • fix(win): fetch a current sd asset (cpu x64), image binary was silently missing (d80b3da)
  • feat(win): device-aware copy + gate Pro to "coming soon" off macOS (c11d8f8)
  • perf/fix(win): GPU (Vulkan) engine + CPU fallback, tighter tool budget, cosine arity (4226098)
  • fix(tools): budget tool schemas to context + surface server errors (B2/B3) (fd11183)
  • fix(win): pin bundled llama-server to b9838 to match the macOS engine (fc112dc)
  • docs(win): add Windows support matrix + test plan/models (10cd1fb)
  • fix(win): fetch-win-binaries crashed on cosmetic listing, failing after copy (5322c17)
  • feat(models): dev control to stop/restart the local model server (5862dcd)
  • fix(llm): guarantee model is loaded before ready + cross-platform orphan cleanup (56f608b)
  • feat(win): cross-platform binary resolution + Windows build pipeline (d58a659)

Full Changelog: v0.0.39-beta.64...v0.0.39-beta.66

Assets 10
Loading