Lemon Squeezy is Poor to agents.
Discry independently scored how well an AI agent can discover and understand the Lemon Squeezy 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 · 17/100Comprehension
55% of score · 81/100What we found
- An agent trying to discover this API programmatically would find almost no machine-readable entry points: robots.txt, llms.txt, llms-full.txt, sitemap.xml, and .well-known/mcp.json all return 404 at the docs domain, so there is no agent-oriented manifest to crawl.
- Once an agent reaches the human docs, comprehension is strong: every endpoint page leads with the HTTP method, path, and a task-oriented one-liner (e.g. 'Creates a unique checkout for a specific variant'), backed by realistic copy-pasteable curl examples with plausible values.
- The developer guide chains operations into genuine end-to-end workflows (create API key, make authenticated requests, create a checkout, handle the post-purchase flow, and persist order/subscription IDs in your database), which is exactly what an agent needs to complete multi-step tasks.
- Naming is rigorously consistent snake_case with ISO 8601 timestamps throughout, and the JSON:API foundation makes request/response shapes predictable, lowering the chance an agent mis-builds a payload.
- No official OpenAPI specification is published; only a third-party (api-evangelist/apis.io) spec is discoverable, so an agent cannot reliably auto-generate a typed client from an authoritative source.
What to change
Prioritized by impact on discoverability. You (or your docs platform) deploy these — Discry never touches your API.
- 01Publish an llms.txt (and ideally llms-full.txt) at docs.lemonsqueezy.com that links the API reference, the developer guide, and the License API, with short section descriptions — this is the single highest-impact discovery fix and currently a hard 404.
- 02Ship an official OpenAPI 3.x specification and link it from the API reference; the API is already JSON:API-structured and stable, so a canonical spec would let agents auto-generate clients instead of relying on unofficial community specs.
- 03Add a robots.txt and sitemap.xml to the docs domain (both 404 today) that explicitly allow AI crawlers and enumerate the documentation pages, so agents can find and index the reference at all.
- 04Add an AGENTS.md to the primary lmsqueezy SDK repos (lemonsqueezy.js, laravel) so coding agents working in those codebases get authoritative setup and usage context.
- 05Consolidate error codes into a single reference with explicit recovery steps (what to do on 401, 422, 429, and license-activation-limit errors) rather than relying on generic HTTP-code descriptions, to give agents self-correction guidance.
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.
Execution characteristics are documented clearly: Bearer-token API-key auth, a JSON:API error envelope with detail/status/title fields, an explicit 300 calls/minute rate limit surfaced via X-Ratelimit-Limit and X-Ratelimit-Remaining headers (429 on overage), and page-based pagination (page[number]/page[size], max 100). Idempotency keys are not mentioned anywhere in the docs.