the backlink api, and the graph underneath it.
if you would rather script backlink research than click through it, there are three layers to work at. the rest api returns a domain's referring domains, or the gap between domains, as json. the mcp server lets an ai assistant call the same tools inside a conversation. underneath both sits the open common crawl graph, which you can process yourself.
the free tier is 15 backlink lookups a month, with a key sent to your email and no card. the posts below cover the free tier, working recipes, and the do-it-yourself route through the raw files.
start here
the free tier, then five working recipes you can paste into a terminal.
crawlgraph now has a free API tier (15 backlink calls a month, no card)
request a key on the docs page and it lands in your inbox. 15 backlink lookups a month on Common Crawl's open web graph, forever free. connect over hosted MCP or run the local npm server.
3 min readprogrammatic backlink research: 5 api recipes
get a free API key, pull one domain's referring sites, filter by authority, compare competitors, feed an automation, and connect hosted MCP, with copyable code for every step.
8 min read
- api documentation. endpoints, authentication, rate limits, and the free key form.
- crawlgraph-mcp on github. the mcp server, run with npx -y crawlgraph-mcp from claude, cursor, or cline.
go deeper
skip the api and process the raw common crawl files yourself.
query the common crawl web graph yourself - backlinks from a 16 gb edge list
three files, no api key, no cluster. the streaming pipeline that pulls one domain's referring domains out of the open graph on a laptop.
8 min read