scout public API

Use scout's public read API to search local LLM recommendations and model metadata. Public endpoints are unauthenticated JSON APIs with CORS enabled.

Base URL: https://llmscout.fit/api

Client guidance: cache responses where practical and stay under roughly 100 requests/minute per IP. Private/admin routes are not public API surface.

GET

`/api/v2/models/search`

Search models by hardware fit, use case, family, text query, and sort order.

curl "https://llmscout.fit/api/v2/models/search?ram=16&usecase=chat&limit=10"

Key query params: ram, vram, usecase, family, sort, q, limit, offset.

GET

`/api/v2/models/:id`

Fetch one model by scout/Hugging Face-style ID.

curl "https://llmscout.fit/api/v2/models/meta-llama--Llama-3.1-8B-Instruct"

GET

`/api/v2/models/leaderboard`

Return top-ranked models by category.

curl "https://llmscout.fit/api/v2/models/leaderboard?category=best&limit=10"

Categories: best, fastest, efficient, trending, community.

GET

`/api/v2/models/stats`

Return collection metadata such as total model count, last update time, family counts, and new-model count.

curl "https://llmscout.fit/api/v2/models/stats"

Private/admin routes

The following routes require bearer tokens and should not be called from browser/client code: POST /api/scrape, GET /api/telemetry, and POST /api/maintenance.

More detail

See the repository documentation for response examples, scoring notes, and legacy endpoints.