◎ Discry Score
mistral ai.com
ai · API
B
0 / 100
DISCOVERY0
COMPREHENSION0
Category leader: 99 (A)
Discry your API →
AI · RANK #24 OF 43

Mistral AI is Good to agents.

Discry independently scored how well an AI agent can discover and understand the Mistral AI API from what’s public — not whether it’s usable. Below: every signal we checked, what’s costing the score, and what to change.

Discry your API — freeView the docs ↗

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 · 81/100
OpenAPI specA machine-readable OpenAPI/Swagger spec agents can parse.Pass
llms.txtAn llms.txt index that points agents to the docs that matter.Pass
llms.txt qualityThe llms.txt is focused, current, and well under the size budget.Pass
llms-full.txtA full-text bundle agents can load in one request.Pass
AGENTS.mdAn AGENTS.md that tells coding agents how to build on the API.Pass
.well-known/mcp.jsonA discoverable MCP manifest at a well-known path.Fail
MCP registryThe API is listed in a public MCP registry.Pass
robots.txt AI directivesrobots.txt allows (or explicitly guides) AI crawlers.Fail
SitemapA sitemap so agents can enumerate the docs surface.Pass

Comprehension

55% of score · 96/100
Task-oriented descriptionsEndpoints described by what they accomplish, not just their shape.Pass
Realistic examplesRunnable, real-world request/response examples.Pass
Multi-step workflowsDocs that chain calls into complete jobs an agent can follow.Pass
Error-recovery guidanceDocumented failure modes and how to recover from them.Partial
Answer-first formatThe answer leads; preamble does not bury it.Pass
Capability boundariesClear limits — what the API can and cannot do.Pass
Naming consistencyConsistent, predictable naming across endpoints.Pass
Heading hierarchyClean heading structure agents can navigate.Pass
Markdown docsDocs available as clean markdown, not JS-rendered HTML only.Pass
Token efficiencyDocs are concise enough to fit an agent context window.Pass

What we found

  • An agent discovering this API lands well: a 14KB API-focused llms.txt with per-page descriptions, a 969KB llms-full.txt, and a clean public OpenAPI 3.1.0 spec (mistralai/platform-docs-public) mean the surface is fully machine-readable without scraping HTML.
  • Every documentation page is published as a parallel .md file (linked from llms.txt), so an agent gets clean markdown for free rather than fighting JS-rendered pages — a major comprehension advantage.
  • Code examples are genuinely agent-usable: realistic model names (codestral-latest, pixtral-12b-2409), real prompts, and python/typescript/curl tabs that are copy-pasteable.
  • The one discovery gap that hurts: robots.txt returns 404 and there is no .well-known/mcp.json — an agent gets no explicit crawl signal and no machine-readable tool manifest at the domain root.
  • Error recovery is the weakest comprehension area: error formats and validation schemas exist, but the docs rarely tell an agent what to DO on a 429/400/401 (e.g. honor Retry-After) — recovery is shown only incidentally via SDK try/except snippets.

What to change

Prioritized by impact on discoverability. You (or your docs platform) deploy these — Discry never touches your API.

  1. 01Add a robots.txt at the docs root that explicitly allows AI crawlers (GPTBot, ClaudeBot, Google-Extended, CCBot) — currently a 404, the single cheapest discovery win.
  2. 02Publish a .well-known/mcp.json (or a documented hosted MCP endpoint) so agents can auto-discover Mistral's tools and auth without third-party registry entries.
  3. 03Add a dedicated error-handling / retry guide that maps the top status codes (429, 400, 401, 422) to explicit recovery steps — e.g. 'on 429, wait the Retry-After interval and back off exponentially'.
  4. 04Document pagination for list endpoints (conversations, files, agents, batch jobs) so an agent can reliably page through large result sets.
  5. 05Surface concrete default rate-limit numbers in the public docs rather than only behind the workspace admin page, so agents can plan request pacing before authenticating.

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.

API key (Bearer token)OAuth2 (for MCP connector clients)✓ Error format documented✓ Rate limits documented✕ Pagination documented✕ Idempotency documented

Auth is a simple Bearer API key (OAuth only appears for MCP connector flows). The OpenAPI spec documents a machine-parseable 422 HTTPValidationError schema, and rate limits are explained conceptually (RPS + tokens/min, tier-based) though exact numbers live behind the workspace admin page. Pagination and idempotency keys are not documented for list/write endpoints.

See your own Discry Score.

Drop your API docs URL. See what an agent sees — in 60 seconds, free.

Discry your API — free