VLM Run launches SystemOne API: typed, calibrated answers in one forward pass
spillai · x · 2026-10-02
VLM Run launched SystemOne, a TypeSafe-compatible API that answers typed questions about text, JSON, or images and returns a calibrated probability for every answer.
- How it works: answers are read off the model in a single forward pass — no decode, no parse — so output is always one of your labels, never off-schema.
- Performance: 44ms p50 over a socket, at a fraction of a cent per read.
- Zero training: engines are open-weight VLMs, so questions can name any concept the model already knows; edit criteria and the next read answers the new question.
- Compatibility: wire-compatible with TypeSafe's Jev API — point the typesafe-sdk client at the gateway and it works unchanged.
Positioned for gating expensive pipelines, routing documents, labeling frames, and sanity-checking uploads. Try via uvx vlmrun gw systemone.
More from coding & agent
- steipete: btrfs CoW is great for worktrees but terrible for sqlite — steipete · 2026-10-02
- Google Mantis: a skills pack for security review with coding agents — udmrzn · 2026-10-02
- NVIDIA and Cedana serve a portfolio of coding models on one 8x B200 node — josh_wills · 2026-10-02
- LukeW teases Intent Mobile: agent coordination at the level of intent — LukeW · 2026-10-02
- LukeW: Dev tools haven't reached their final form after Terminal UI and chat clients — LukeW · 2026-10-02
- Dev Benchmarks 10+ Coding Agent Workflows: Astra Planning + Sol 6.1 Beats Pricier Combos — kevinkern · 2026-10-02