BACKLNK

BACKLNK MCP server

Free technical SEO, AEO, LLMO and GEO checks for AI agents: check a page, fix it, then verify the fix.
Evidence for every finding, most important first. Remote MCP server, free API key.

Get a free API key · Connect your agent · Tools · Checks · REST API

---

BACKLNK is a remote Model Context Protocol server. It gives Claude Code, Cursor, VS Code and other MCP clients the same evidence-based checks that run on backlnk.com: meta tags and social previews, page structure, structured data, AI crawler access in robots.txt, redirects and headers, sitemaps, indexability, hreflang, internal links, duplicate content, images, llms.txt and page speed.

Every finding comes with its evidence (the URL, the exact excerpt and where it was found) and a recommendation, most important first. Your agent fixes the code, and after you deploy, verify_fixes re-runs the same check and tells you what is fixed, what is new, what is still open and what could not be re-checked.

Endpoint	https://backlnk.com/mcp (Streamable HTTP, stateless)
Authentication	Free API key: Authorization: Bearer bk_... (or X-API-Key: bk_...)
Price	Free. 100 checks per key per 24 hours, plus a fair-use hourly limit
Official MCP Registry	com.backlnk/backlnk
Documentation	backlnk.com/developers

This repository holds the documentation, client configuration examples and the registry entry. The server itself runs at backlnk.com; there is nothing to install.

Get a key

1. Open backlnk.com/developers and enter your e-mail address.
2. Click the link in the confirmation e-mail, then press the button on the page it opens.
3. Your key (bk_...) is shown once. Copy it and keep it somewhere safe.

Requesting a new key later replaces the old one.

Connect your agent

Claude Code

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

Cursor

Add to ~/.cursor/mcp.json (all projects) or .cursor/mcp.json (one project). See examples/cursor-mcp.json.

{
  "mcpServers": {
    "backlnk": {
      "url": "https://backlnk.com/mcp",
      "headers": { "Authorization": "Bearer YOUR_KEY" }
    }
  }
}

VS Code

Add to .vscode/mcp.json. VS Code asks for the key once and stores it securely, so the file can be committed. See examples/vscode-mcp.json.

{
  "inputs": [{ "type": "promptString", "id": "backlnk-api-key", "description": "BACKLNK API key", "password": true }],
  "servers": {
    "backlnk": {
      "type": "http",
      "url": "https://backlnk.com/mcp",
      "headers": { "Authorization": "Bearer ${input:backlnk-api-key}" }
    }
  }
}

Other clients

Any MCP client that supports remote servers over Streamable HTTP works: point it at https://backlnk.com/mcp and send the key in the Authorization header (Bearer bk_...). Clients that can only forward a plain header can use X-API-Key: bk_... instead.

Try it

Ask your agent:

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.

The loop:

1. Check: run_checks on the page's public URL. You get the issues and warnings, most important first, each with evidence and a recommendation.
2. Fix: the agent edits your templates, metadata, structured data or server configuration.
3. Verify: once the change is deployed to that URL, verify_fixes with each audit id. It re-runs the check fresh and compares: fixed, new, still open, not re-checked.

A finding counts as fixed only when the re-check assessed the same rule and found no problem. If the re-check could not assess it, BACKLNK says so instead of claiming a fix.

Tools

Tool	What it does	Key
list_checks	Lists every check (id, what it looks at, its scope and limits).	yes
run_checks	Runs checks on one public URL and waits up to about 40 seconds. Returns per check its state, summary, counts and the issues and warnings to fix, plus audit ids. Default set: meta-tags, page-structure, ai-crawler-access, redirects-headers, entity-schema.	yes
get_report	The compact report of an earlier check. detail: "actions" (default) for issues and warnings, "all" for every finding.	yes
verify_fixes	Re-runs the same check fresh on the same URL and options and returns the comparison with the original run.	yes
ai_crawler_policy	Builds robots.txt lines that allow or block AI crawlers from BACKLNK's reviewed crawler list, with what each line does and warnings. Presets: block-training, block-ai, allow-all, or decide per purpose and per crawler.	yes
usage	How many checks the key has used and has left today and this hour.	yes

Listing the tools (tools/list) works without a key; calling them needs one.

Arguments

• run_checks: url (required), checks (check ids, optional), mode (Source, Rendered or Auto), fresh (re-run even if the same check ran very recently).
• get_report: audit_id (required), detail (actions or all).
• verify_fixes: audit_id (required).
• ai_crawler_policy: preset, or purposes (allow or block per purpose) and crawlers (per token, e.g. { "GPTBot": "allow" }).

Status NotAssessed means a check could not determine the result: it is not a pass. Only Issue and Warning findings need action.

Checks

Id	Check	Scope
meta-tags	Meta & social preview	One page and its share image.
page-structure	Structure & readability	One page.
entity-schema	Schema & entities	One page; up to 15 sameAs destinations are checked for reachability.
ai-crawler-access	AI crawler access	robots.txt of the host plus one page's meta robots and X-Robots-Tag.
redirects-headers	Redirects & headers	The URL you enter plus its http:// and www / non-www variants.
image-seo	Image SEO	One page; up to 60 image files measured.
page-speed	Page speed	One page, two lab runs (mobile and desktop).
indexability	Indexability	Up to 50 URLs on one website, plus its robots.txt and sitemap.
sitemap-hygiene	Sitemap audit	Up to 20 sitemap files and 50,000 URLs; up to 40 destinations sampled.
internal-links	Internal links	Same host, up to 40 pages and depth 4.
hreflang	Hreflang	Same website, up to 40 pages reached through hreflang alternates and links.
duplicate-content	Duplicate & thin content	Same website, up to 40 pages.
llms-txt-generate	llms.txt draft	Reads up to 25 of your pages.
llms-txt-validate	llms.txt validation	Checks up to 60 listed links.

The live list is always at https://backlnk.com/api/v1/checks. Two website checks that use a hosted language model (Answer snippets, Intent & focus) run only on backlnk.com, not through the API or the MCP server.

REST API

The same checks without MCP, for scripts and CI. Base URL https://backlnk.com/api/v1. See examples/rest-api.sh.

# 1. Start a check (returns 202 with the audit id)
curl -X POST https://backlnk.com/api/v1/audits \
  -H "Authorization: Bearer YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{"url":"https://example.com/","checks":["meta-tags","page-structure"]}'

# 2. Read the result (poll until "state" is Completed, Partial or Failed)
curl https://backlnk.com/api/v1/audits/AUDIT_ID -H "Authorization: Bearer YOUR_KEY"

# 3. After fixing the page: re-check and compare with the first run
curl -X POST https://backlnk.com/api/v1/audits/AUDIT_ID/verify -H "Authorization: Bearer YOUR_KEY"

Method	Path	Key
GET	/checks	no	The checks you can run.
POST	/audits	yes	Start checks on one URL: { "url", "checks": [ids], "mode"? , "options"? }. An optional Idempotency-Key header makes retries safe.
GET	/audits/{id}	yes	State, summary, scores, coverage and findings, most important first (?detail=actions for issues and warnings only).
POST	/audits/{id}/verify	yes	Re-run fresh and compare with the original.
GET	/audits/{id}/export?format=json|csv	yes	The full report (JSON) or the findings (CSV).
GET	/usage	yes	Today's and this hour's usage of the key.
POST	/ai-crawler-policy	no	robots.txt lines for AI crawlers from a preset or your choices.

Errors are JSON, { "error": { "code", "message" } }, with the matching HTTP status: 401 for a missing or revoked key, 429 with Retry-After when a limit is reached.

Limits and fair use

• 100 checks per key in any 24 hours; results BACKLNK reuses from a very recent identical check are free.
• A fair-use hourly limit per key, and limits per website so no site is crawled too hard.
• BACKLNK only checks public pages. Private, local and internal addresses are refused.
• GET /api/v1/usage and the usage tool show what is left.

Privacy

BACKLNK keeps your e-mail address with your key and deletes it if you ask (contact). Reports are private to the key that ran them and are never published or indexed.

Support

Questions, bugs or ideas: open an issue or write to us at backlnk.com/contact.

License

The documentation and examples in this repository are available under the MIT License. The BACKLNK service is free to use within its limits.