snakereachAPI reference
The HTTP API.
One base URL, six endpoints, and one error shape. Everything under /v1 needs an API key, request one. You can try every endpoint without a key in the playground.
https://api.snakereach.comEverything here
Start
Endpoints
- POST /v1/scrapeAcquire one URL: fetch, extract everything, return one normalized document.
- POST /v1/discoverEvery link on a page, canonicalized and classified as internal or external.
- POST /v1/sitemapFind and expand a site's sitemaps, or its feeds, into a flat URL list.
- POST /v1/crawlCrawl a whole site as a background job, then poll it for results.
Reference
- The Document schemaEvery field snakereach returns: markdown, metadata, links, images, tables, records, trace.
- The performance tracePer-stage timings for every acquisition, fetch, parse, content, metadata, links, media, records, combine.
- Rate limitsPer-IP request limits, concurrency ceilings, and crawl budgets on the snakereach API.
- ErrorsEvery status the snakereach API returns, in RFC 9457 problem+json, with how to fix each.
The shape of it
Every acquisition produces the same Document: markdown first, metadata complete, with the links, images and tables the page contained and a trace of how long each stage took.
Nothing is scored or ranked. The extraction runs every harvester on every page and merges the results with fixed precedence, so the same input gives the same output.