API Integration Guide

The TeamBench API enables programmatic access to content review capabilities, allowing you to integrate AI-powered quality scoring directly into your existing content management systems, publishing workflows, and custom applications.

The API follows REST conventions and communicates via HTTPS using JSON for request and response bodies. All API endpoints are available under the base URL for your TeamBench instance. Authentication uses token-based auth -- you include your API token in the Authorization header of each request as a Bearer token.

To obtain an API token, log in to TeamBench and navigate to your account settings. Generate a new API token, which will be displayed once. Copy and store it securely -- it will not be shown again. This token authenticates all your API requests and has the same permissions as your user account within your organization.

The primary integration use case is submitting content for review programmatically. Send a POST request to the review endpoint with the content text, the reviewer ID, and optionally the model to use. The API returns a JSON response containing the overall score, per-criterion scores and feedback, and metadata about the review. Your application can then process these scores -- display them in your CMS, enforce quality gates in your publishing pipeline, or store them for analytics.

Reviewer management endpoints allow you to list available reviewers, retrieve reviewer details including criteria and weights, and create or update reviewers via the API. This enables you to manage your review infrastructure programmatically, which is useful for organizations that need to provision reviewers across multiple content workflows.

Review history endpoints let you retrieve past reviews, filter by date range or reviewer, and export scores for analysis. This data can feed dashboards, reporting systems, or quality tracking databases that your organization already uses.

Error handling follows HTTP conventions. Successful requests return 200-series status codes with the expected response body. Client errors (invalid input, missing parameters, authentication failures) return 400-series codes with descriptive error messages. Server errors return 500-series codes and should be retried after a brief delay.

Rate limiting applies to API requests to ensure fair usage and system stability. Standard rate limits are generous for typical integration use cases. If you anticipate high-volume API usage, contact the TeamBench team to discuss rate limit adjustments for your account.

When designing your integration, consider implementing retry logic with exponential backoff for transient failures, caching reviewer configurations that do not change frequently, and processing reviews asynchronously if your application does not require immediate results. These patterns improve reliability and efficiency for production integrations.