Product
Developers
The original developer direction was Openclaw: smart model routing powered by peer-reviewed LLM Council deliberations. Ask any question and get back the best model for the job.
Openclaw API
Openclaw was designed around a simple idea: send a query and return the best-performing LLM for that topic, based on historical council performance.
Under the hood, LLM Council asks multiple frontier models the same question, has them review one another, and records which models consistently win for different topics.
Routing logic
The product learns from council runs: selected models, peer-review results, winners, consensus, disagreement, timing, and task type.
That data lets LLM Council route future work more intelligently instead of treating every prompt as a generic chat completion.
- Topic-aware model selection
- Council performance history
- Cost-aware routing
- Traceable decisions for why a model or panel was chosen
Current surface
Public API access is not the main staging surface yet. The hosted LLM Council app is the production path first: authenticated sessions, working context, document workflows, council runs, exports, and billing-managed tiers.