BACKLNK

npx add-mcp 'https://backlnk.com/mcp'

Free API key: MCP server and REST API

BACKLNK inside your AI agent

Let Claude Code, Cursor or VS Code run BACKLNK's technical SEO and AI-visibility checks on a page, fix what they find in your code, then verify the fix. Every finding comes with the evidence behind it.

Check, fix, verify

Your agent gets findings it can act on, and proof that the change worked, without leaving your editor.

1. Step 1

   Check

   The agent runs BACKLNK's checks on a page and gets back the issues and warnings, most important first, each with evidence and a recommendation.
2. Step 2

   Fix

   It edits your templates, metadata, structured data or server configuration in your codebase, the way you review any other change.
3. Step 3

   Verify

   It re-runs the same checks fresh after the change is deployed and gets a comparison: what is fixed, what is new and what is still open.

Try this prompt

Use BACKLNK to check https://example.com/pricing. Fix the issues it finds in this repository, most important first. When I've deployed, verify the fixes.

Run this once in your terminal. Claude Code then offers the BACKLNK tools in every session; check with /mcp.

Terminal

claude mcp add --transport http backlnk https://backlnk.com/mcp --header "Authorization: Bearer YOUR_KEY"

Endpoint reference

Base URL https://backlnk.com/api/v1. Send Authorization: Bearer YOUR_KEY and JSON bodies. Responses are JSON and never cached; errors are {"error":{"code","message"}} with the matching HTTP status (401 for a missing or revoked key, 429 with Retry-After when a limit is reached).

• GET /checks The checks you can run: id, tool, what they look at and their limits.
• POST /audits key required Start checks on one URL. Body: { "url", "checks": [ids], "mode"?: "Source" | "Rendered" | "Auto", "options"?: { "maxPages", "maxDepth", "declaredPurpose", "audience" } }. Optional Idempotency-Key header makes retries safe. Returns 202 with one audit per check.
• GET /audits/{id} key required State, summary, scores, coverage and findings, most important first (?detail=actions for issues and warnings only). Verifications include a comparison.
• POST /audits/{id}/verify key required Re-runs the same check fresh on the same address and options. The new audit compares itself with the original: fixed, new, still open, not re-checked.
• GET /audits/{id}/export?format=json|csv key required Download the full report (JSON) or the findings (CSV).
• GET /usage key required How much of today's and this hour's quota the key has used.
• POST /ai-crawler-policy robots.txt lines for AI crawlers from a preset or your choices. Body: { "preset" } or { "purposes", "crawlers" }.

Example: GET /audits/{id} (shortened)

Findings come most important first: issues, then warnings, by severity. Only issues and warnings have a severity other than None.

Response

{
  "id": "0f8fad5b-d9cb-469f-a165-70867728950e",
  "check": "meta-tags",
  "url": "https://example.com/pricing",
  "state": "Completed",
  "summary": "…",
  "counts": { "issues": 1, "warnings": 2, "informational": 3, "passed": 14, "notAssessed": 0, "notApplicable": 1 },
  "findings": [
    {
      "ruleId": "…",
      "status": "Issue",
      "severity": "High",
      "title": "…",
      "explanation": "…",
      "recommendation": "…",
      "evidence": [{ "url": "https://example.com/pricing", "excerpt": "…", "location": "line 12:5" }],
      "documentation": "https://…"
    }
  ],
  "links": { "self": "…", "verify": "…", "exportJson": "…", "exportCsv": "…" }
}

Check ids

Use these in checks. GET /checks returns them with their inputs, scope and limits.

• llms-txt-generate llms.txt draft
• llms-txt-validate llms.txt validation
• entity-schema Schema & entities
• sitemap-hygiene Sitemap audit
• internal-links Internal links
• page-structure Structure & readability
• ai-crawler-access AI crawler access
• meta-tags Meta & social preview
• redirects-headers Redirects & headers
• indexability Indexability
• hreflang Hreflang
• duplicate-content Duplicate & thin content
• image-seo Image SEO
• page-speed Page speed

Answer snippets and Intent & focus use a hosted language model, so they run only on backlnk.com, not through the API or the MCP server.

Frequently asked questions

Is the API really free?

Yes. The key, the REST API and the MCP server are free: no cost, no card and no account. Each key has a daily and an hourly limit so the checks stay fast and polite for everyone; your agent can read its remaining quota at any time.

Why do I have to confirm my e-mail?

So that each key belongs to a real, reachable address. We e-mail a link first; the key is shown once, on the page that opens after you press the confirm button. If you didn't ask for a key, ignore the e-mail and nothing happens.

I lost my key. Can you send it again?

No. The key is shown once, when you confirm, and can't be retrieved again. Request a new key with the same address: once you confirm it, the old key stops working.

Which pages can I check?

Public web pages. BACKLNK fetches them politely, within fixed limits per check; private, local and internal addresses are refused. A change your agent makes locally is only visible to the checks once it is deployed to a public address.

What do you do with my e-mail address?

We use it to send your key and keep it with the key, and we delete it if you ask. See our privacy page for details.

Prefer a browser? Every check also runs on its own on the tools page, and the full audit e-mails all of them as one PDF.