Clootrack is Not Agent-Ready to agents.
Discry independently scored how well an AI agent can discover and understand the Clootrack API from what’s public — not whether it’s usable. Below: every signal we checked, what’s costing the score, and what to change.
SCORED UNDER RUBRIC 1.2 · A full re-launch under Discry Score 2.5 — a new behavioral instrument, not comparable to these scores — is in progress.
Discovery
45% of score · 38/100Comprehension
55% of score · 13/100What we found
- An agent looking for a way to call Clootrack programmatically would find no public REST API and no developer documentation portal — the marketing site, third-party listings (GetApp), and search all confirm there is no openly documented API surface to reason over.
- Clootrack's entire agent strategy is MCP-first: a managed MCP server (mcp.clootrack.com) is promoted on /mcp and listed in the Glama registry, but the connector is currently 'Unhealthy'/credential-gated and its actual configuration lives behind a DocSend link and a sales demo — an autonomous agent cannot self-serve onboarding or discover its tools.
- A well-behaved crawler is welcomed: robots.txt blocks no AI bots and an llms.txt exists — but the llms.txt is a 284KB site-wide dump of ~776 links (397 of them blog posts) with essentially zero API-relevant content, so it actively wastes an agent's context rather than orienting it.
- There are no endpoint descriptions, code examples, authentication docs, error catalogs, rate limits, or multi-step workflow guides anywhere public — an agent would learn Clootrack's capabilities and limits only by contacting sales, not by reading docs.
- The strongest agent-readiness signal Clootrack has is its MCP FAQ, which does explain access control, PII handling, and supported data sources in plain language — useful capability context, but marketing-grade and no substitute for a tool/parameter reference.
What to change
Prioritized by impact on discoverability. You (or your docs platform) deploy these — Discry never touches your API.
- 01Publish a public MCP tool reference (tool names, inputs, outputs, auth flow) instead of gating configuration behind a DocSend link — an agent-first product should let agents discover and understand its MCP tools without a sales call.
- 02Serve a .well-known/mcp.json declaring the MCP server endpoint and tool declarations so agents can auto-discover the server from the domain rather than only via the Glama registry.
- 03Replace the 284KB blog-heavy llms.txt with a right-sized, MCP/API-focused llms.txt (<50KB) that links to the MCP tool reference, capability boundaries, and auth setup, so an agent can orient in a few thousand tokens.
- 04Document capability boundaries and execution characteristics explicitly — supported datasets, rate/volume limits, error formats, and recovery guidance — so agents can self-correct instead of failing blind.
- 05Fix the malformed sitemap.xml (URLs are concatenated into a single entry) and add any MCP/config documentation pages to it so discovery crawlers can index the agent-relevant surface.
Execution coverage · INFORMATIONAL, UNSCORED
Whether an agent can actually complete a call and recover from errors is the deeper Audit layer — documented here, but not part of the Discry Score.
Clootrack exposes no public REST API and publishes no developer/API documentation, OpenAPI spec, authentication guide, or endpoint reference. Its only agent-facing surface is a managed MCP server (mcp.clootrack.com), described on a marketing page (/mcp) with configuration gated behind a DocSend link and a sales demo. Execution characteristics (error format, rate limits, pagination, idempotency) are undocumented publicly; access is governed by platform role-based access control and is enterprise/credential-gated.