Skip to main content
This section lists every public Data Hub REST endpoint by resource. Each endpoint page covers method and path, auth, parameters, real response samples and field notes, possible errors, and the matching client-library method. This page only holds conventions shared by every endpoint.

Base URL

The /v1 contract is additive only: new fields and endpoints may appear over time, but published fields keep their names and meaning. Ignore unknown fields when you integrate, and do not depend on field order.

Authentication

Authenticated endpoints use standard Bearer auth. Put the Data Hub API key in the Authorization header:
Create an API key in the Octoparse account center. Access tokens from web login or OAuth go in the same place. Each endpoint page labels its auth class:
The Data Hub API only accepts Authorization: Bearer. It does not accept credentials in URL query parameters. This is not the same convention as the Octoparse scraper MCP in the top navigation, which uses an x-api-key header. Do not mix them. Treat an API key like an account credential: never commit it to a repo, shared config, or public screenshot.

Response shape

Successful responses wrap the payload in data. Error responses wrap the payload in error. The two never appear together.
A few endpoints return non-JSON content (raw Markdown, CSV / JSONL text, or a zip binary). Those pages call that out explicitly.

Errors

Each endpoint page lists codes unique to that endpoint. These appear on most endpoints: Returning the same 404 for “missing” and “invisible” is intentional: private apps, other people’s runs, and other people’s datasets never leak existence through the error code.

Pagination

List endpoints use offset / limit query parameters and return a pagination object:
total is the filtered total. When has_more is true, add count to offset and continue. Per-endpoint limit caps are documented on each page.

Time parameters

Endpoints that take a time range (run list, billing aggregation, publisher analytics) accept ISO 8601 timestamps. Prefer an explicit offset such as 2026-09-01T00:00:00+08:00. Unix epoch seconds are not accepted.

App references

Anywhere an app is identified, both forms work: Point-to-point shared apps do not appear in market search. List them with the shared-with-me filter on the search endpoint. For scheduled jobs and long-lived automation, store app_id.

Run statuses

Endpoint groups

Discover Data Apps

Search, detail, versions, README, specs, and translations.

Runs and results

Start, poll, list, cancel, read records, and debug traces.

Datasets

Persistent containers for run results and retention flags.

Account and billing

Lifetime spend and billing aggregates by period and dimension.

Secrets

Upstream credentials that app authors store for their own apps.

Publishing and operations

Draft → Build → Release, ops switches, sharing, translations, contract tools, and publisher analytics.