Rapidata Developer Portal

Rapidata provides human feedback at scale — crowd-sourced labeling, model evaluation, ranking, and preference data. Integrate with the Python SDK or directly against the OAuth 2.0 secured REST API.

Get started

• Quick Start — install, authenticate, create your first job
• Install the SDK: pip install rapidata
• Create API credentials (client ID & secret)

Authenticate as an agent

The API uses OAuth 2.0 (OpenID Connect), client-credentials grant:

1. Create a client ID & secret.
2. Set RAPIDATA_CLIENT_ID and RAPIDATA_CLIENT_SECRET in the environment.
3. Construct RapidataClient() — it fetches and refreshes a bearer token automatically.
4. Without the SDK: POST https://auth.rapidata.ai/connect/token with grant_type=client_credentials, then send the token as Authorization: Bearer ….

Full details: Authentication.

Reference

• OpenAPI specification (/openapi.json)
• API reference — the RapidataClient and its managers
• Authentication — OAuth 2.0 / OpenID Connect, scopes, token endpoint
• OpenID Connect discovery document

Guides & examples

• Documentation home
• Custom audiences, results, early stopping
• Examples — classify, compare, locate, draw, rank, free text

For AI agents

• Use Rapidata from your AI agent — official coding-agent skill
• llms.txt · llms-full.txt

Source

• GitHub · PyPI