Liminal 0.1.0 — first public preview (beta)

v0.1.0 is the first tagged public preview: a stability milestone with CI on every push to main, a reproducible eval sandbox (npm run eval:sandbox), inference latency mode, a desktop managed inference catalog for model selection, and sidecar multi-chat support. This is beta — RC and install GA are still ahead on the roadmap.

Shipped in this slice#

• CI workflow — .github/workflows/ci.yml runs build, typecheck, and test on every push to main
• Eval sandbox framework — packages/eval/fixtures/ sandbox scenarios; npm run eval:sandbox for reproducible agent tests
• Inference latency mode — AGENT_LATENCY_MODE patches for performance-optimized inference paths
• Tool circuit — tool_circuit.ts for parallel tool execution optimization
• Desktop managed inference catalog — Flutter UI for managed inference model selection
• Sidecar multi-chat — chat_registry.ts for multi-session support
• Release checklist — docs/RELEASE-CHECKLIST.md documents the release process

What you can run today#

# Download Liminal Desktop (Windows / macOS / Linux)
# https://www.vireondynamics.com/liminal/get-started#desktop-app

# Run eval sandbox (no API key required)
npm run eval:sandbox -w packages/eval

Roadmap#

Pre-release history#