Skip to documentation

External REST API

Read and write the products, accounts, evidence, decisions, Initiatives, Studies, and workspace context that connected systems need. The API uses JSON over HTTPS and workspace API keys. Vendor connections such as Zoom, Jira, and Google Drive are covered in the integration guides; this reference is for HTTP paths, methods, and payloads.

  1. 1. Create a scoped key

    Choose only the resources and actions the integration needs.

  2. 2. Resolve stable IDs

    List products, teams, or records before writing relationships.

  3. 3. Follow returned state

    Use pagination, retry headers, and job status routes as returned.

API resources

Choose the resource that matches the record or workflow you need. Each reference page keeps its endpoints, request contract, responses, and examples together.

TasksFind sprint work, update tasks and attach pull requests.8 endpointsSprintsRead team sprints and resolve the current sprint.3 endpointsSignal evidenceRead exact evidence and mark delivered work as addressed.3 endpointsAnalyticsCompute bounded workspace aggregates without downloading raw records.1 endpointSignalsImport and manage transcripts, tickets, reviews, and research evidence.6 endpointsAccountsResolve customer organizations, people, activity, and external identities.15 endpointsProductsRead and maintain product context, features, personas, and questions.18 endpointsInitiativesRead Initiatives, then maintain lifecycle fields, team ownership, and generated documents.6 endpointsStudiesInspect Study formats, accepted context, privacy-bounded findings, and freshness.7 endpointsTeam HuddlesRetrieve daily team history, saved snapshots, artifacts, and conversation transcripts.3 endpointsFeedbackReconcile rating events and calculate transparent fixed-window rollups.2 endpointsInsightsRead and maintain evidence-backed customer and market learnings.6 endpointsOpportunitiesMaintain evidence-backed product problems and opportunity records.5 endpointsIdeasMaintain solution candidates, questions, votes, and product links.10 endpointsTaxonomiesRead, publish, and apply the workspace classification system.18 endpointsOKRsRead and maintain objectives and key results.5 endpointsContext UnitsMaintain reusable, structured workspace context.5 endpointsContext DocumentsMaintain long-form workspace context documents.7 endpoints

For agents and tooling

A reference agents can retrieve directly

The Markdown reference is generated from the endpoint objects used by these pages, so agents get the same operations, parameters, scopes, responses, and examples as people do.

Open the complete Markdown index

Authentication

All API requests must include an Authorization header with your API key as a bearer token. You can manage your API keys in the Settings > API Keys section of your Zentrik workspace.

Zentrik workspace Settings showing the API keys area before any keys exist.

Create and manage keys from Settings → API keys in your workspace.

bash
curl -H "Authorization: Bearer YOUR_API_KEY" \
  https://zentrik.ai/api/external/v1/insights
Modal for creating an API key with scope checkboxes.

Each key can be limited to the API scopes your automation needs.

Grant only the scopes your integration needs. Taxonomy automation usestaxonomies:read,taxonomies:write,taxonomies:publish,classification-runs:read,classification-runs:write, andclassification-queue:read depending on whether the client only previews definitions or also publishes and refreshes classifications.

Workspace and team access. An API key is a workspace-wide integration principal. It can access every record allowed by its scopes, including Initiatives assigned to any team in the workspace. Selecting a team in the app does not restrict the key. MCP connections use the signed-in person’s current team memberships instead.

Keep your API keys secret. Do not share them or commit them to public repositories. If a key is compromised, revoke it immediately in the dashboard.

Requests, pagination, and limits

Send JSON over HTTPS. Paginated catalog endpoints return the requested items in the response body and describe the page in response headers.

Pagination

Start with offset=0. Advance by X-Limit while X-Has-More is true. The maximum page size is 100.

Rate limits

The standard policy is 10 requests per second, 100 per minute, and 1,000 per hour for each API key. Contract-specific limits can differ. Read the returned RateLimit headers instead of hard-coding a delay.

Retries

A 429 response includes Retry-After. Wait for that many seconds. Retry transient failures with bounded exponential backoff, but fix validation and permission errors before retrying.

http
X-Total-Count: 42
X-Limit: 20
X-Offset: 0
X-Has-More: true
RateLimit-Policy: 10;w=1, 100;w=60, 1000;w=3600
RateLimit-Limit: 10
RateLimit-Remaining: 9
RateLimit-Reset: 1

Error responses

Errors use the HTTP status code as the primary contract and include a concise message. Rate-limit errors also include reasonCode, retryable, and retryAfterSeconds.

Create a workspaceAsk supportDocumentation