Free · Signed · No account required
Download OpenAlchemy Engine
Run llama.cpp, whisper.cpp, and VibeVoice locally. Serve those runtimes back to OpenAlchemy and earn credits when other users dispatch jobs to your GPU.
Windows 10 / 11
Bundled with CUDA-12 llama.cpp + whisper.cpp + VibeVoice runtime packs. ~30 second install.
- Version
- v0.3.0
- Installer size
- 510 MB
- Format
- .msi
Every release is minisign-signed and verified by the updater. View update manifest · Public key 2635A8BB661F3404.
What's new in this release
Full release notes →- Single-channel runtime panel
Stable / Beta tabs are gone. One catalog, one list. No more ghost rows, no "Switch to Beta" prompts, no "Show experimental" toggle.
- Active runtimes is truly global
The top status row no longer changes when you scroll through pack lists. Display names come from a cross-channel pack metadata cache.
- Version-scoped pack URLs
Every runtime pack is fetched from a per-version path so re-publishing never silently overwrites a binary that older clients still reference.
Available under the terms of use. Telemetry and data handling follow the privacy policy.