Conventions
Shared API conventions
Base URL & versioning
BASH
1https://api.zenode.ai/v1The major version lives in the path. Additive changes — new optional fields, new endpoints — ship
within v1 and are not breaking. Anything breaking would ship under a new major version.
Request & response format
- Requests and responses are
application/json; charset=utf-8. - Field names are
snake_case; timestamps are ISO-8601 UTC. - Every response includes a
request_id(also in theX-Request-Idheader). - Authenticated responses include
X-RateLimit-Limit,X-RateLimit-Remaining, andX-RateLimit-Reset— see Rate limits & usage.
Methods
POSTfor lookups, matching, and search — they take rich JSON bodies (and have no URL-length limits).GETfor fetching a specific resource by its slug (part detail, manufacturer detail, jobs).
Result lists are bounded, not paginated
Match and catalog-search endpoints return a ranked, bounded list controlled by limit. There is
no cursor or offset — ask for a larger limit if you need more candidates.
| Parameter | Type | Required | Description |
|---|---|---|---|
/parts/match | 10 / 50 | Optional | Ranked by confidence |
/parts/match/batch | 5 / 50 per line | Optional | Up to 250 lines per request |
/parts/search | 10 / 25 | Optional | Compact PartSummary per hit |
/parts/discover | 5 / 10 | Optional | AI-ranked picks |
The one cursor-paginated collection is the AI job list — those are your own jobs, not catalog data.
Usage billing
REST has two paid classes: full part records and AI tokens. Matching, summary catalog search, taxonomy lookups, filter facets, job management, and usage reads are free and abuse-tracked.
See Usage billing for the full endpoint table and how each class is counted.
Next
- Usage billing — which endpoints cost what
- Object reference — the structure of every response object
- Part matching