ollaya
Ollaya serves a specific kind of model: decision models that read a state, such as a message, ticket or any JSON, plus typed questions (choice, score or yes/no) and return calibrated probabilities in a single forward pass. They never generate text, so answers come back in milliseconds. A support message can be scored for intent, urgency, frustration, refund request and churn risk in one call.
The workflow mirrors Ollama: ollaya serve runs a daemon, and run, pull, list, ps, show, rm, cp, stop and create behave as you would expect. Its /v1/systemone, /v1/decisions and /v1/models endpoints are wire-identical to TypeSafe's hosted Jev API, so the TypeSafe SDK and existing Jev clients work by changing the base URL. Ollaya publishes only small ONNX graphs; weights load from each author's Hugging Face repository, pinned to a commit and checked by sha256, and GGUF models run on llama.cpp.
For agents, ollaya mcp exposes the models to Claude Code, Cursor and other MCP clients, and a skill teaches agents when to use them. The README's recommended model is a 4B-class model that wants an NVIDIA GPU; on CPU the smaller laya router models are the suggested start.
Features
Why choose it
Trade-offs
Maintenance check
- Last commit
- 2026-11-06
- Commits, last 90 days
- 100+
- Releases, last 90 days
- 27
- Latest release
- v0.18.0 (2026-10-25)
- Open issues
- 11
- Recent issues with at least one comment
- 1/3
Oldest week on the left; the latest four weeks are orange. Source: GitHub commit activity.
Status rule: archived on GitHub = Archived; last commit within 90 days = Active; 91 to 365 days = Slowing; older = Stale. All figures come from the GitHub API on the date shown.
Most discussed open issues
- #80 Decima 2.3: decima-guard (new), decima-agent 2.3, and the choice-window fix you found9 comments · 2026-10-18
- #79 Support Drex (nace-ai/drex-v2.0): Qwen3.6-35B-A3B MoE with a Kev pointer head, GGUF2 comments · 2026-10-17
- #165 Support hs-decide-1-flash1 comments · 2026-11-04
- #170 Feature Request: Gate model pulls0 comments · 2026-11-10
- #168 please support adding remote open ai compatible endpoint0 comments · 2026-11-09
Issue titles are quoted from GitHub. Open the link to check whether a problem still applies to the version you use.
Compatibility
Quick start
Use cases
Alternatives
Comments
No comments yet. Be the first!