# Rapidata

> Rapidata provides human feedback at scale: crowd-sourced labeling, model
> evaluation, ranking, and preference data (RLHF-style) collected from real
> people. The supported way to access it programmatically is the Rapidata
> Python SDK, documented on this site (docs.rapidata.ai).

Integrate via the Python SDK (`pip install rapidata`). The API is OAuth 2.0 /
OpenID Connect secured; authenticate with a client ID/secret created at
https://app.rapidata.ai/settings/tokens, or via interactive browser login on
first run.

- Developer portal: https://docs.rapidata.ai/developers/
- OpenAPI specification: https://docs.rapidata.ai/openapi.json
- Authentication (OAuth 2.0): https://docs.rapidata.ai/latest/authentication/
- OpenID Connect discovery: https://auth.rapidata.ai/.well-known/openid-configuration

## When to use Rapidata

Reach for Rapidata when a task needs **real human judgment at scale** rather than
a model's own output:

- Evaluate or compare model outputs (image, video, audio, text) with human preferences — pairwise A/B comparisons, ranking, leaderboards.
- Build RLHF / preference datasets and reward-model training data.
- Annotate or label data: classification, locating, bounding/drawing, transcription, free text, select-words.
- Collect targeted human feedback by country, language, or demographic, with built-in quality control and confidence-based early stopping.

It is **not** the right tool for fully automated labeling with no human in the
loop, or for sub-second real-time inference.

How an agent calls it: use the Python SDK (`pip install rapidata`, see Quick
Start) or the OAuth 2.0 REST API (see Authentication). Coding agents can install
the official skill (see AI agents) to have integrations written for them.

## Guides

- [Overview](https://docs.rapidata.ai/latest/starting_page/): what Rapidata does and its core concepts
- [Quick Start](https://docs.rapidata.ai/latest/quickstart/): install, authenticate, and create your first order
- [Custom Audiences](https://docs.rapidata.ai/latest/audiences/): target responses by country, language, and qualification
- [Signals](https://docs.rapidata.ai/latest/signals/)
- [Parameter Reference](https://docs.rapidata.ai/latest/job_definition_parameters/)
- [Understanding Results](https://docs.rapidata.ai/latest/understanding_the_results/)
- [Early Stopping](https://docs.rapidata.ai/latest/confidence_stopping/)
- [Instruction Design](https://docs.rapidata.ai/latest/human_prompting/)
- [Error Handling](https://docs.rapidata.ai/latest/error_handling/)
- [Logging & Config](https://docs.rapidata.ai/latest/config/)

## Examples

- [Classification](https://docs.rapidata.ai/latest/examples/classify_job/)
- [Comparison](https://docs.rapidata.ai/latest/examples/compare_job/)
- [Locate](https://docs.rapidata.ai/latest/examples/locate_job/)
- [Draw](https://docs.rapidata.ai/latest/examples/draw_job/)
- [Select Words](https://docs.rapidata.ai/latest/examples/select_words_job/)
- [Free Text](https://docs.rapidata.ai/latest/examples/free_text_job/)
- [Ranking](https://docs.rapidata.ai/latest/examples/ranking_job/)

## Model ranking & benchmarks

- [Getting Started](https://docs.rapidata.ai/latest/mri/)
- [Advanced](https://docs.rapidata.ai/latest/mri_advanced/)

## API & AI agents

- [Developer portal](https://docs.rapidata.ai/developers/): all developer resources in one place
- [OpenAPI specification](https://docs.rapidata.ai/openapi.json): the full API surface (OAuth 2.0 / OpenID Connect)
- [Authentication](https://docs.rapidata.ai/latest/authentication/): OAuth 2.0 client-credentials flow, scopes, token endpoint
- [API reference](https://docs.rapidata.ai/latest/api/): the `RapidataClient` class and its managers
- [Use Rapidata from your AI agent](https://docs.rapidata.ai/latest/ai_agents/): an official skill that teaches coding agents (Claude Code, Cursor, Copilot, and others) to write Rapidata integrations

## Access

- Install: `pip install rapidata`
- API tokens: https://app.rapidata.ai/settings/tokens
- Source: https://github.com/RapidataAI/rapidata-python-sdk
- PyPI: https://pypi.org/project/rapidata/

## Optional

- [llms-full.txt](https://docs.rapidata.ai/llms-full.txt): the full documentation concatenated into a single file