Wayfinder Router: deterministic routing of queries between local and hosted LLM
The Wayfinder Router is a deterministic routing system that sends queries to either local or cloud models based on the prompt's structure and wording. It decides in microseconds, runs offline, and never calls another model to make the call, reducing latency, cost, and randomness. The system is designed to be self-hosted and calibrated on the user's own data, allowing for a score and recommendation to be provided for each prompt.
- ▪Wayfinder Router uses a deterministic approach to route queries between local and cloud models.
- ▪The system reads the structure and wording of a prompt to decide where to send it, eliminating the need for model calls and API keys.
- ▪Wayfinder Router can be calibrated on the user's own data and provides a score and recommendation for each prompt.
Opening excerpt (first ~120 words) tap to expand
Deterministic prompt-complexity routing — send each prompt to your local or cloud model, offline, with no model call to decide. Quickstart · Benchmark · How it compares · Explainer · Changelog No model callto decide the route Deterministicand fully offline Calibrateon your own data Bring your own keyself-hosted Wayfinder looks at a prompt's structure (length, headings, lists, code) and its wording (proofs, math, hard constraints), then tells you whether to send it to your small local model or your big cloud one. It decides in microseconds, runs offline, and never calls another model to make the call: no API key, no network, no model call to decide. You get a score and a recommendation, and what you do with it is up to you.
…
Excerpt limited to ~120 words for fair-use compliance. The full article is at GitHub.