# juanmc2005/diart

A python package to build AI-powered real-time audio applications

Repository: https://github.com/juanmc2005/diart
Canonical: https://ross.abutalabs.com/products/diart
Homepage: https://diart.readthedocs.io
Language: Python
License: MIT
License Family: permissive
Topics: speaker-diarization, streaming-audio, real-time, speaker-embedding, deep-learning, transcription, voice-activity-detection
Last push: 2026-06-19T22:36:53+00:00

## Health v2 (maintenance only)
Score: 62/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 88, release rhythm 8, longevity 100
- inputs: {"age_days": 1850, "days_push": 75, "days_rel": 567, "gap_med": null, "n_releases_24m": 1}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 2022, forks 165 (observed 2026-08-28T04:06:06.425077+00:00)

## What it is
Diart is a Python framework for building AI-powered real-time audio applications, best known for state-of-the-art streaming speaker diarization. It combines speaker segmentation and embedding models with incremental clustering, and supports custom pipelines, hyper-parameter tuning, benchmarking, and serving over websockets.

## Use cases
- identify who is speaking in a live audio stream
- real-time speaker diarization for meetings or calls
- stream voice activity detection over websockets
- build custom real-time speech AI pipelines
- benchmark and tune online diarization hyper-parameters
- live captioning with speaker labels

## When to choose
- you need low-latency, streaming speaker diarization in Python
- you want to prototype real-time audio AI pipelines with pre-trained models
- you need to serve audio AI models over websockets

## When to avoid
- you only need offline (batch) diarization of recorded files
- you need production transcription, which is still marked as coming soon
- you work outside the Python/PyTorch ecosystem

## Facets
- artifact type: framework
- maturity: active
- function: audio-processing, speech-recognition, machine-learning, deep-learning, streaming, benchmarking
- domain: speech-processing, machine-learning, artificial-intelligence
- platform: python, cross-platform
- tags: speaker-diarization, streaming-audio, voice-activity-detection, speaker-embedding, websockets, real-time-speech, audio, real-time

## Member repositories
- juanmc2005/diart (main) score 62

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:06:06.425077+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T03:00:06.319704+00:00, confidence not recorded.
  - readme: https://github.com/juanmc2005/diart (fetched 2026-08-28T04:06:06.425077+00:00, sha 988efe9a3134)
  - registry_pypi: https://pypi.org/pypi/diart/json (fetched 2026-08-29T10:40:36.111616+00:00, sha 5facf5ee0182)
- Data as of 2026-08-30T08:39:29.467469+00:00.