# pyannoteAI

> pyannoteAI is a speaker intelligence platform that turns conversational audio into structured speaker metadata: speaker diarization (who spoke when), speaker identification (matching voices to known identities via voiceprints), speaker separation (splitting overlapping voices into clean individual streams, coming later), and STT Orchestration (a specialized pipeline and reconciliation logic that produces highly accurate speaker-attributed transcription). Live-1 provides real-time streaming diarization. Models run through the hosted API, on-premises, or on-device.

## Agent Instructions

- This file indexes the Markdown sources agents should prefer over crawling HTML pages.
- Request pages on www.pyannote.ai with the `Accept: text/markdown` header, or append the `md` query parameter with no value, for example `https://www.pyannote.ai/pricing?md` to get the Markdown version of the page.
- To get the Markdown version of any page on https://docs.pyannote.ai, append `.md` to the URL, for example `https://docs.pyannote.ai/quickstart.md`.
- Full URL list is at https://www.pyannote.ai/sitemap.xml.

## Terminology

- Speaker diarization: the process of identifying and tracking individual speakers in an audio recording.
- Speaker identification: the process of determining the identity of a speaker.
- Voiceprint: biometric signature of a speaker's voice, used for speaker identification.
- Diarization error rate (DER): the diarization accuracy metric. Missed speech, false alarm speech and speaker confusion, summed and divided by total speech duration.
- Voice activity detection (VAD): the process of identifying speech regions in an audio recording.

## Website

- [Home](https://www.pyannote.ai/?md)
- [Pricing](https://www.pyannote.ai/pricing?md): plans, rates, and deployment options including on-premises and self-hosting.
- [Speaker platform](https://www.pyannote.ai/speaker-platform?md): the full framework and how it fits into an existing pipeline.
- [Precision-3](https://www.pyannote.ai/precision-3?md): the flagship speaker diarization model.
- [Streaming](https://www.pyannote.ai/streaming?md): real-time diarization with Live-1.
- [STT Orchestration](https://www.pyannote.ai/stt-orchestration?md): speaker-attributed transcription with any STT engine.
- [Enterprise](https://www.pyannote.ai/enterprise?md): on-premises, self-hosted, and sovereign deployment.
- [Benchmarks](https://www.pyannote.ai/benchmark?md): diarization error rate comparisons.
- [Solutions](https://www.pyannote.ai/solutions?md): AI voice agents, contact center, healthcare and AI scribing.
- [Use cases](https://www.pyannote.ai/usecases?md)
- [Customer stories](https://www.pyannote.ai/customer-story?md)
- [Changelog](https://www.pyannote.ai/changelog?md): list of changes and updates.
- [Blog](https://www.pyannote.ai/blog?md): index of all blog posts.
- [About us](https://www.pyannote.ai/about-us?md)
- [Contact sales](https://www.pyannote.ai/contact-us?md)
- [Book a call](https://www.pyannote.ai/book-a-call?md)

## Solutions

- [AI voice agents](https://www.pyannote.ai/solutions/ai-voice-agents?md)
- [Contact center](https://www.pyannote.ai/solutions/contact-center?md)
- [Healthcare and AI scribing](https://www.pyannote.ai/solutions/healthcare-ai-scribing?md)
- [Media and dubbing](https://www.pyannote.ai/solutions/media-dubbing?md)
- [Transcription and indexing](https://www.pyannote.ai/solutions/transcription-indexing?md)
- [Voice model training](https://www.pyannote.ai/solutions/voice-model-training?md)

## Documentation

- [Documentation index, all pages](https://docs.pyannote.ai/llms.txt)
- [Introduction](https://docs.pyannote.ai/introduction.md)
- [Quickstart guide](https://docs.pyannote.ai/quickstart.md)
- [Models](https://docs.pyannote.ai/models.md)
- [Feature overview](https://docs.pyannote.ai/features.md)
- [Data retention](https://docs.pyannote.ai/data-retention.md): compliance, data region, data lifecycle and deletion policies.
- [Billing](https://docs.pyannote.ai/administration/billing.md): billing, spend limits and subscription management.

## Tutorials

- [Diarize an audio file](https://docs.pyannote.ai/tutorials/how-to-diarize-audio.md)
- [Diarize live audio](https://docs.pyannote.ai/tutorials/streaming-real-time.md)
- [STT Orchestration](https://docs.pyannote.ai/tutorials/speech-to-text-diarization.md): speaker-attributed transcripts in a single API call.
- [Identify speakers with voiceprints](https://docs.pyannote.ai/tutorials/identification-with-voiceprints.md)
- [Configure the number of speakers](https://docs.pyannote.ai/tutorials/speaker-configuration.md)
- [Use the API playground](https://docs.pyannote.ai/tutorials/api-playground.md)

## Open Source

- [pyannote.audio](https://github.com/pyannote/pyannote-audio/blob/main/README.md): the open-source speaker diarization toolkit.
- [Community-1 model card](https://huggingface.co/pyannote/speaker-diarization-community-1)
- [Hugging Face organization](https://huggingface.co/pyannote)

## Optional

- [Dashboard and API playground](https://dashboard.pyannote.ai)
- [Trust and compliance center](https://trust.pyannote.ai)
- [Status and uptime](https://status.pyannote.ai)
- [Terms of use](https://www.pyannote.ai/terms-of-use?md)
- [Privacy policy](https://www.pyannote.ai/privacy-policy?md)
- [Support](mailto:support@pyannote.ai)
- [Discord community](https://discord.gg/vux8UH9QmV)