SeeklynOpen the app

API & MCP docs

Seeklyn's deterministic scoring engine is a public API, and every AI Search Intelligence report is a JSON endpoint. No black-box model — every number comes from auditable rubrics or raw counts. Use it in curl, CI pipelines, or agent tools.

Start here — no signup, no key

Score any headline in one curl call. Free, rate-limited per IP, no account required.

curl "https://api.seeklyn.com/api/v1/analyze?title=Your+headline"

No signup. No key. Free.

Remote MCP — works today

Seeklyn runs a public streamable-HTTP MCP server. Paste the URL below into Claude.ai custom connectors, Cursor's MCP settings, or ChatGPT developer mode — no API key, no installation.

https://api.seeklyn.com/mcp

Once connected, two tools are available to your agent:

  • analyze_passage — score a passage of text for AI citability: BM25 retrieval simulation, extractability lint, per-chunk findings.
  • audit_url — fetch a URL and return its website citability score.

Transport: streamable HTTP, MCP protocol 2024-11-05. No metered LLM calls — responses arrive in milliseconds.

GitHub Action — citability lint in CI

Add citability checks to pull requests. The action runs a static Go binary — no API calls, no keys, completes in milliseconds — and posts inline PR annotations for every passage that scores below your threshold.

# .github/workflows/citability.yml  — coming soon (public via seeklyn-oss)
name: Citability
on: pull_request
jobs:
  lint:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v4
      # Public action ships with the seeklyn-oss mirror — not yet consumable externally
      - uses: seeklyn-oss/citability-lint@v1
        with:
          files: 'content/**/*.md docs/**/*.md'
          threshold: '65'

Coming soon. The action runs today inside Seeklyn's own (private) repo; a public, externally-usable version ships with the upcoming seeklyn-oss mirror. Until then, use the keyless API or the remote MCP above in CI.

REST endpoints

All endpoints live under https://api.seeklyn.com. Analysis endpoints are free and keyless; metered endpoints (cited checks, history) require an API key.

POST /api/v1/analyze

Headline scoring with full JSON body. Accepts title, h1 (optional — powers title↔H1 mismatch detection), mode, and a context object with keyword. Returns the same HeadlineResult as the GET form — score, band, CTR/AEO sub-scores, rewrite risk, word balance, suggestions, and more.

POST /api/v1/analyze-content

Passage and website citability analysis. Send {title, content} for article text or {url, type:"website"} to analyze a live URL. Returns the Citation Funnel (retrieval → citation → overall), BM25 retrieval simulation, per-chunk scores and flags, query fan-out coverage, and deterministic before/after fixes.

Authenticate with an API key for higher limits and result history:

Authorization: Bearer sk_…
# or
X-Seeklyn-Key: sk_…

Issue and revoke keys from the account settings page. Full schema: openapi.yaml.

AI Search Intelligence data

Every report in the dashboard's AI Search Intelligence suite is a JSON endpoint under /api/v1/projects/{id}/ai/…, authenticated with your browser session or an API key (keys inherit your plan and workspace access). The project id demo returns the public, read-only sample dataset with no key — the same sample the dashboard shows before your own prompts have run.

curl "https://api.seeklyn.com/api/v1/projects/demo/ai/summary?days=30"
  • Observe — ai/answers · ai/answers/{runId}/{engine} · ai/model-changes
  • Measure — ai/summary · ai/volatility · ai/markets
  • Setup — ai/prompts · ai/segments · ai/prompts/discover · integrations
  • Compete — ai/competitive · ai/competitive/gaps · ai/sources
  • Diagnose — ai/opportunities · ai/narrative · ai/accuracy · ai/content-map · ai/audit
  • Impact — ai/rankings · ai/search-console · ai/generative-report · ai/backlinks · ai/business

Every report takes the dashboard's global filters: engine, country and days. Each response carries its provenance (your data, sample, or simulated) and raw counts, never a proprietary score. For bulk data use ai/export?dataset=answers|citations|mentions|prompts&format=csv|json (up to 50,000 rows). Two endpoints need no project: /api/v1/ai/engines (the AI engines Seeklyn tracks) and /api/v1/ai/methodology (every formula).

stdio MCP & CLI binary

A local stdio MCP server (seeklyn-mcp) and a citability lint CLI (seeklyn-lint) will ship as go install targets and cross-compiled binaries via the upcoming seeklyn-oss public mirror. No npm package exists yet. For zero-install use today, use the remote MCP URL above.