> For clean Markdown of any page, append .md to the page URL.
> For a complete documentation index, see https://dev.hume.ai/audio-decisions/docs/llms.txt.
> For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://dev.hume.ai/_mcp/server.

# Documentation

## Docs

- [Audio Decisions API](https://dev.hume.ai/audio-decisions/docs/overview.md): Ask questions about audio in plain language, and get answers that estimate how human raters would respond.
- [For coding agents](https://dev.hume.ai/audio-decisions/docs/for-coding-agents.md): What a coding agent needs to call the Audio Decisions API correctly, and how to point an agent at these docs.
- [Use cases](https://dev.hume.ai/audio-decisions/docs/use-cases.md): Example questions to send the Audio Decisions API, grouped by what they ask about the audio.
- [Authentication](https://dev.hume.ai/audio-decisions/docs/authentication.md): Authenticate requests with an API key.
- [Python quickstart](https://dev.hume.ai/audio-decisions/docs/quickstart/python.md): Install the Python SDK, ask four questions about a recording and read the answers.
- [TypeScript quickstart](https://dev.hume.ai/audio-decisions/docs/quickstart/typescript.md): Install the TypeScript SDK, ask four questions about a recording and read the answers.
- [Preparing your audio](https://dev.hume.ai/audio-decisions/docs/preparing-your-audio.md): How to choose whose voice and how much time each request covers, and what to keep in the audio you send.
- [Writing questions](https://dev.hume.ai/audio-decisions/docs/writing-questions.md): How to write questions that get useful answers, and which questions to avoid.
- [Interpreting answers](https://dev.hume.ai/audio-decisions/docs/interpreting-answers.md): How to turn each answer into a decision, and how to set thresholds from audio you have labeled.
- [Evaluating many files](https://dev.hume.ai/audio-decisions/docs/evaluating-many-files.md): Prepare a set of recordings, send one request per file within the rate limit, handle failures and find the runs that errored.
- [Primitives](https://dev.hume.ai/audio-decisions/docs/primitives.md): The four question types, what each asks and what the numbers in each answer describe.
- [Runs](https://dev.hume.ai/audio-decisions/docs/runs.md): Fetch the status and request of past decisions requests.
- [Support](https://dev.hume.ai/audio-decisions/docs/support.md): How to get help, report a security issue and reach the sales, legal and press teams.
- [Limits](https://dev.hume.ai/audio-decisions/docs/limits.md): The limits on audio, questions, context and request rate, and the error a request over each one returns.
- [Errors](https://dev.hume.ai/audio-decisions/docs/errors.md): The error format and codes of the Audio Decisions endpoints.
- [Billing](https://dev.hume.ai/audio-decisions/docs/billing.md): How Audio Decisions API usage is billed, how to estimate it before you send audio and where to check it.