Free tier is enough to build with. Pro removes the caps when you're running it in production.
Both plans query the same proprietary editorial-authority index. Wondering how the data compares to Ahrefs? See the FAQ below.
Both plans include full REST + MCP access. No dashboard required, though one exists if you want it.
| Feature | Free | Pro |
|---|---|---|
| Lookups per day | 10 / IP | Unlimited |
| REST API access | Yes | Yes |
| MCP server access | Yes | Yes |
| Linker list rows returned | Up to 25 | Up to 100 (full ranked set) |
| Signup required | No | Yes |
| Support | Community / docs | Priority email |
An editorial-authority score (a domain_rating-style score from 0-100), referring domain counts, and ranked lists of linking domains with their own authority and outbound-link counts — computed from BacklinkMCP's proprietary web-scale link index. It's the same shape whether you call REST or MCP.
It's a proprietary index, refreshed periodically and optimized for fast, stable queries. verify_link_live fetches each linking domain's homepage (not inner pages) to confirm a live href today. See the docs for the exact field shapes.
Ahrefs runs its own continuous web crawler and starts at $129/mo, dashboard-first. BacklinkMCP is $0-19/mo, API/MCP-first, and built specifically so agents and scripts can query it directly — see the full Ahrefs alternative comparison.
Free tier is 10 lookups/day per IP address, tracked per calendar day. Pro removes the daily cap entirely. Both REST and MCP tool calls count against the same per-IP or per-key limit — see rate limits in the docs.
Self-serve checkout is not live. Email hello@backlinkmcp.com to request an operator-issued Pro key. There is no card charge path on this site today, so there are no Pro charges to refund.
Use MCP if your client is Claude Desktop, Claude Code, Cursor, or another MCP-aware agent — it's zero-code, add the server URL and the thirteen tools show up. Use REST if you're calling from a script, backend service, or anything that isn't an MCP client. Both hit the identical cached engine and count against the same rate limit.