Programmatic access to your AI visibility data — scans, citations, competitors, census-weighted prompt volumes, proactive signals and SEO rankings. Available on Standard plans and above.
Create an API key in Settings → API Keys. Keys look like rk_live_… and are shown once at creation. Pass them as a Bearer token:
curl -H "Authorization: Bearer rk_live_YOUR_KEY" \ "https://app.rankie.ai/api/v1/scans?limit=5"
| Endpoint | Description | Query params |
|---|---|---|
GET /api/v1/scans | List completed scans | limit (≤50), clientId |
GET /api/v1/scans/:id | Scan summary with per-model breakdown | clientId |
GET /api/v1/citations | Citations for a scan (5-category typed) | scanId (required), clientId |
GET /api/v1/competitors | Competitor mention shares for a scan | scanId (required), clientId |
GET /api/v1/prompt-intelligence | Census-weighted AI prompt volume clusters | scanId (required), clientId |
GET /api/v1/recommendations | Prioritized action recommendations for a scan | scanId (required), clientId |
GET /api/v1/strategy | Full Hemingway growth strategy (roadmap, quick wins, content plan) | scanId (optional — latest if omitted), clientId |
GET /api/v1/signals | Proactive change signals (score drops, rank moves) | days (≤30), clientId |
GET /api/v1/seo-rankings | Tracked keyword position history | days (≤365), clientId (required) |
Agency accounts: pass clientId to scope any endpoint to one of your clients. Omitting it returns your own brand's data only — never a client's. Rate limit: 60 requests/minute. Errors are JSON: {"error": "..."} with 401 / 403 / 404 / 429 status codes.
A hosted MCP endpoint exposes the same data as 16 tools for Claude, Cursor and any MCP client — including get_prompt_intelligence (census-weighted AI query volumes, unique to Rankie) and start_scan (requires a key with write scope). Authentication uses the same API keys.
Claude Code
claude mcp add rankie https://app.rankie.ai/api/mcp \ --transport http \ --header "Authorization: Bearer rk_live_YOUR_KEY"
Cursor / other clients (mcp.json)
{
"mcpServers": {
"rankie": {
"url": "https://app.rankie.ai/api/mcp",
"headers": { "Authorization": "Bearer rk_live_YOUR_KEY" }
}
}
}Tools: list_scans, get_scan_summary, get_citations, get_competitors, get_prompt_intelligence, get_recommendations, get_strategy, get_signals, get_seo_rankings, start_scan, and Executor Agent tools: list_agent_tasks, get_agent_task,create_task, approve_task (requires write + tasks:approve — approvals go through the same function the panel uses, never bypassing the review queue), reject_task / reject_tasks (bulk, up to 50 with one reason; require tasks:reject — a reject-only key cannot approve, publish or write; tasks:approve keys can reject too), get_impact_report. REST mirrors:GET/POST /api/v1/agent/tasks, POST /api/v1/agent/tasks/:id/approve|reject, POST /api/v1/agent/tasks/reject (bulk, per-id results). Agency accounts pass client_id per tool call.