Etherpad API

{
  "service_type": "platform",
  "base_url": "https://etherpad.org",
  "auth_method": "api_key",
  "auth_config": {
    "api_key_param": "apikey",
    "notes": "Each Etherpad instance generates an API key accessible via the admin UI (Admin > API). Self-hosted instances use this key for all API requests. The hosted etherpad.org site is the project's homepage, not a shared instance; API access targets whatever instance a user runs or uses."
  },
  "endpoints": [],
  "pricing_model": {
    "type": "free",
    "details": {
      "license": "Apache License 2.0",
      "note": "Open source software, free to self-host. Costs depend on the operator's own infrastructure. Third-party managed hosting may charge separately."
    }
  },
  "rate_limits": {
    "note": "No global rate limits enforced by the project; limits depend on the specific instance deployment and its reverse proxy configuration."
  },
  "capabilities": [
    "Real-time collaborative text editing for thousands of simultaneous users",
    "Self-hosted deployment on your own server",
    "Full pad data export (HTML, plain text, Etherpad format)",
    "REST-like HTTP API for programmatic pad management",
    "Programmatic pad creation, deletion, and copying",
    "Read/write pad content via API",
    "Author and session management (createAuthor, createSession)",
    "Pad access control and read-only mode",
    "Plugin ecosystem for extending functionality",
    "Optional database backends (DirtyDB, MySQL, PostgreSQL)",
    "Cookie-based sessions for embedding and integration",
    "Chat within pads",
    "Revision history and timeslider",
    "Import/export of documents"
  ],
  "raw_analysis": "Etherpad is an open-source, self-hosted real-time collaborative editor, originally developed by AppJet and now maintained by the Etherpad Foundation. It is a mature project (lineage dating to ~2008, actively maintained) aimed at teams, organizations, and developers who need a privacy-respecting, self-controlled alternative to Google Docs or similar services. It scales to thousands of concurrent users per instance depending on deployment resources.\n\nThe REST-like API is exposed at the path /api/1/ on each Etherpad instance (not on etherpad.org itself, which serves as the project's marketing/documentation site). Authentication uses an api_key query parameter, with the key configured per-instance in the admin settings (settings.json or the admin web UI). API responses are structured JSON (versions 1.2.13+ default to JSON; older versions support JSONP).\n\nKey API groups include pad operations (createPad, deletePad, copyPad, getRevisionsCount, getLastEdited, getReadOnlyID, setPublicStatus, getPublicStatus, setPassword, listAuthorsOfPad, getPadID), text content operations (getText, setText, appendText, getHTML, setHTML, getAttributePool, getRevisionChangeset, getAuthorId), and author/session management (createAuthor, createAuthorIfNotExistsFor, listPadsOfAuthor, createSession, getSessionInfo, deleteSession).\n\nThere are no official rate limits; throughput is governed by the instance's server resources and any reverse-proxy throttling imposed by the operator. Pricing is effectively free: the software is Apache-2.0 licensed, and users pay only for their own hosting. Several commercial managed Etherpad providers exist (e.g., hosted services) but the core platform itself is FOSS.\n\nIntegration surface: a plugin system allows custom hooks and extensions, and pads can be embedded via iframes with cookie/session-based authentication. It is commonly integrated into learning management systems, wikis, and internal collaboration suites. Maturity is high, with a stable API version (1.x) and broad language client library coverage (e.g., Python, Node.js, PHP, Java)."
}

1/3 tests passed

{
  "overall": 68,
  "dimensions": {
    "token_efficiency": 8.0,
    "first_try_success": 6.0,
    "response_parseability": 6.5,
    "error_clarity": 6.5,
    "doc_quality": 7.5,
    "auth_simplicity": 7.0,
    "latency": 10.0,
    "consistency": 7.5
  },
  "pricing_normalized": {
    "model": "free_open_source",
    "license": "Apache-2.0",
    "self_hosted_required": true,
    "managed_hosting_available_third_party": true,
    "effective_cost_for_agents": "infrastructure + ops burden, no SaaS signup",
    "note": "No paywalled API tiers; cost is purely self-hosting overhead."
  },
  "issues": [
    "No robots.txt (404) — agents/crawlers have no crawl guidance.",
    "No llms.txt (404) — no machine-readable entry point for LLM agents to discover API/docs.",
    "No security headers detected on homepage — weak signal for agent-trust/posture checks.",
    "Self-hosted only; agents cannot onboard a user to a managed endpoint without operator setup (Node, DB, reverse proxy).",
    "Auth is cookie/session + author tokens (createAuthor/createSession) — workable but not a clean modern API-key/Bearer flow; friction for agent-driven automation.",
    "REST-like API is not fully REST; documentation inconsistencies historically noted by community.",
    "HTML-first site; no native JSON endpoint for platform metadata, forcing HTML parsing for discovery."
  ],
  "recommendations": [
    "Publish /llms.txt and /robots.txt to make the platform machine-discoverable by agents.",
    "Ship a public OpenAPI/JSON schema for the HTTP API to improve response_parseability and first_try_success.",
    "Document a canonical auth flow with a single API key or token, plus a 'quickstart in <5 min' guide targeting automated operators.",
    "Expose a managed hosted tier (or partner-managed endpoints) so agents can onboard users without deployment.",
    "Add standard security headers (HSTS, CSP, X-Content-Type-Options) on the marketing/docs site.",
    "Provide example Docker Compose / one-click deploy templates to cut first_try_success friction.",
    "Standardize error responses (JSON error codes) across pad/content endpoints.",
    "Add a machine-readable pricing/deployment page (self-host vs managed) so agents can compare options concisely."
  ]
}