# Couchbase — API Agent Score

> Score: 91/100 (Grade: A+) | Domain: couchbase.com | Rubric: 1.0.0 | Checked: October 1, 2026

Couchbase scored 91/100 (A+), classified "ai-ready".

[View full report](https://www.postman.com/ai/ai-ready-apis/company/couchbase)
[All organizations](https://www.postman.com/ai/ai-ready-apis/llms.txt)

## Summary

- Overall: 91/100 (A+)
- Classification: ai-ready

## Category Scores

- API Design: 98/100 (A+)
- Developer Experience: 74/100 (B)
- Agent Discovery: 92/100 (A+)
- Agent Understanding: 95/100 (A+)
- Agent Usability: 90/100 (A+)

## Check Results

### API Design

- [pass] Machine-readable, versioned contract — The API declares version "v4.0" and uses url-based versioning, so an agent can pin to a stable contract. Investigated: spec 100%, docs 100%.
- [pass] Auth declared & discoverable — Authentication is required across the API (declared globally or on every operation), so an agent knows credentials are needed. Investigated: spec 100%, docs 100%.
- [warn] Schema coverage & depth — 13 of 220 public signatures use loose types (anyTypeRatio=0.059). Investigated: sdk 93%, spec 50%.
- [pass] Security & governance hygiene — Code is formatter-clean with zero linter errors and zero style warnings. Investigated: sdk 100%, spec 0%, wellknown 0%.
- [pass] Example coverage — 90% of parameters and responses (1769 of 1961) include example values, grounding an agent in real payload shapes. Investigated: spec 100%, sdk 100%.

### Developer Experience

- [pass] Self-service developer portal — Self-service signup available at https://docs.couchbase.com/ai/get-started/intro.html, with a free tier or sandbox documented — an agent can onboard without contacting sales. Investigated: docs 100%.
- [pass] Description completeness — 94% of operations and parameters carry substantive descriptions (250 of 285 ops, 177 of 177 params), the main semantic cue an agent relies on. Investigated: spec 100%, sdk 71%.
- [warn] Quickstart present — Quickstart page is reachable at https://docs.couchbase.com/home/index.html (24 variants scanned) but has no runnable code sample in its HTML or .md variant, so there's nothing to copy and run. Investigated: docs 50%.
- [fail] Changelog published — No changelog link or mention appears in the spec, so an agent can't tell what changed before something breaks. Investigated: spec 0%, docs 0%.
- [skip] Code samples in docs — No surface produced evidence for this capability in this run.

### Agent Discovery

- [pass] Docs reachable, not hard auth-gated — All 1 pages are publicly accessible. Investigated: docs 100%.
- [pass] Registry & SDK presence — Indexed on Context7 (websites/couchbase, 69631 snippets). Investigated: docs 100%, sdk 100%, cli 0%.
- [warn] llms.txt present, valid & comprehensive — llms.txt covers 0/1 sitemap doc pages (0%); 1 missing — Missing: https://docs.couchbase.com/home/index.html. Investigated: docs 79%, sdk 0%.
- [warn] Crawlable / AEO — Sitemap has 51 entries but no lastmod values, so crawlers can't tell what changed recently. Investigated: wellknown 75%.

### Agent Understanding

- [pass] Machine-readable errors (RFC 9457) — Documents 12 distinct 4xx/5xx error response codes, so an agent can branch on specific failures instead of guessing. Investigated: docs 100%, sdk 100%, spec 50%.
- [pass] Operation purpose clarity — 99% of operations (281 of 285) have both a clear summary and a descriptive name, so an agent can pick the right endpoint. Investigated: spec 100%, sdk 100%.
- [pass] Agent-navigable, token-efficient docs — All 1 pages contain server-rendered content. Investigated: docs 100%.
- [warn] Agent instructions file (AGENTS.md) — No agents.md or skill.md context file found (nice-to-have for agent operation). Investigated: sdk 50%, wellknown 0%.
- [skip] Docs structured data — No surface produced evidence for this capability in this run.
- [skip] Description consistency across surfaces — Only 1 surface description(s) with ≥6 tokens available; need at least 2 to compare.

### Agent Usability

- [pass] Idempotency documented — Built-in retry machinery is present (grep-derived). Investigated: sdk 100%, spec 0%, docs 0%.
- [pass] Pagination documented & consistent — Pagination parameters are present on the 46 list endpoint(s), so an agent can page through results. Investigated: spec 100%, docs 100%.
- [pass] Sandbox separation — Sandbox environments are exposed (sandbox). Investigated: sdk 100%, spec 0%, docs 0%.
- [warn] Rate-limit signaling — Only 1 rate-limit response header(s) are documented, so an agent has limited signal for backing off before being throttled. Investigated: spec 50%, docs 50%.
- [pass] Runnable collection with test scripts — README is present and includes a runnable quickstart. Investigated: sdk 100%, platform 0%.

## Executive Summary

Couchbase's external API program is performing strongly across agent discoverability, agent understanding, API design, and operational reliability — partners and their agents can find, parse, and run your APIs with confidence. The one meaningful gap is in partner trust around change management: there is no published changelog, meaning partners and their AI agents have no way to anticipate breaking changes before they cause integration failures. Closing this single gap should be the immediate priority to protect partner program stability and reduce support escalation risk.