◎ Discry Score
openai.com
ai · API
A
0 / 100
DISCOVERY0
COMPREHENSION0
Category leader: 99 (A)
Discry your API →
AI · RANK #11 OF 43

OpenAI is Agent-Ready to agents.

Discry independently scored how well an AI agent can discover and understand the OpenAI 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 · 90/100
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.Pass
SitemapA sitemap so agents can enumerate the docs surface.Partial

Comprehension

55% of score · 100/100
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.Pass
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

  • OpenAI achieves near-perfect scores across both dimensions: the llms.txt on developers.openai.com is API-focused and right-sized, llms-full.txt provides a complete 2.9MB markdown export, and AGENTS.md exists in multiple repos — all three new v1.1 checks pass
  • Multi-step workflow documentation is exemplary: end-to-end guides for Codex workflows, ChatKit integration, and Agent Builder provide step-by-step instructions for chaining operations
  • Error recovery guidance is actionable: error codes page includes specific resolution steps (rate limit handling, retry logic, common parameter mistakes) that an agent can follow programmatically
  • The only discovery gap is the missing .well-known/mcp.json — despite OpenAI being a major MCP ecosystem participant, automated MCP discovery is not available
  • Platform.openai.com sitemap.xml contains only /tokenizer — the developers.openai.com domain has fuller coverage but the primary docs domain is underserved

What to change

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

  1. 01Add .well-known/mcp.json to developers.openai.com with tool declarations for core API capabilities and auth configuration
  2. 02Host or mirror the OpenAPI spec on the primary docs domain instead of relying on the Stainless-hosted URL
  3. 03Expand platform.openai.com sitemap.xml to include API documentation pages or redirect to developers.openai.com
  4. 04Document idempotency key support (if available) to help agents safely retry failed requests

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_keyoauth2✓ Error format documented✓ Rate limits documented✓ Pagination documented✕ Idempotency documented

API key auth via Bearer token with organization/project headers. Detailed error codes with HTTP status codes, Python SDK error types, and resolution steps. Rate limits with 5-tier system, per-model limits, and x-ratelimit-* headers. Cursor-based pagination. No idempotency key support documented.

See your own Discry Score.

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

Discry your API — free