Adyen is Good to agents.
Discry independently scored how well an AI agent can discover and understand the Adyen 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 · 64/100Comprehension
55% of score · 87/100What we found
- Adyen's llms-full.txt is a 28MB concatenation of every public markdown page — the most comprehensive single-document agent resource found in payments, though its size (>50KB) means it requires RAG chunking rather than single-prompt consumption.
- AGENTS.md files exist in multiple Adyen GitHub repos (adyen-flutter, adyen-node-api-library, and all other SDK repos) — a clear investment in AI-assisted development.
- Every documentation page is available as native markdown by appending .md to the URL (e.g., docs.adyen.com/online-payments.md), making the entire docs site agent-consumable.
- No robots.txt on docs.adyen.com (404) — agents won't find explicit guidance on crawling permissions for the documentation subdomain.
- The llms.txt is well-structured with 136 sections covering all Adyen products but is more of a site-wide index than an API-focused guide.
What to change
Prioritized by impact on discoverability. You (or your docs platform) deploy these — Discry never touches your API.
- 01Add robots.txt to docs.adyen.com explicitly allowing AI bots and referencing the llms.txt file.
- 02Create a trimmed, API-focused llms.txt variant (under 50KB) that covers only REST API integration patterns rather than all product documentation.
- 03Add sitemap.xml to docs.adyen.com listing API documentation pages specifically.
- 04Add .well-known/mcp.json to formalize the MCP discovery path for agents.
- 05Document specific rate limits publicly — current docs reference limits but don't provide numbers.
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.
Well-documented authentication via API key header, structured error responses, cursor-based pagination in balance platform APIs. Rate limits not publicly documented with specific numbers. Idempotency supported via Idempotency-Key header.