# studio-dots-ai/dots.tts

Repository: https://github.com/studio-dots-ai/dots.tts
Canonical: https://ross.abutalabs.com/products/dotstts
Language: Python
License: Apache-2.0
License Family: permissive
Last push: 2026-08-17T03:46:50+00:00

## Health v2 (maintenance only)
Score: 79/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 98, release rhythm 97, longevity 6
- inputs: {"age_days": 90, "days_push": 16, "days_rel": 21, "gap_med": 0, "n_releases_24m": 4}
- flags: young
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1275, forks 130 (observed 2026-08-28T04:04:12.851904+00:00)

## What it is
dots.tts is a 2B-parameter fully continuous, end-to-end autoregressive text-to-speech system, distributed as a Python library with pretrained checkpoints on Hugging Face. It combines a semantic encoder, an LLM backbone, and a flow-matching acoustic head over a 48 kHz AudioVAE, achieving open-source state-of-the-art results on Seed-TTS-Eval and multilingual speaker-similarity benchmarks, with a companion dots.tts.edit model for instruction-controlled speech editing.

## Use cases
- convert text into natural sounding speech from python
- clone a voice from a short reference audio sample
- generate multilingual speech across 24 languages
- synthesize emotional and expressive speech
- edit or rewrite spoken words in an existing recording via text instructions
- run high-fidelity 48 kHz speech synthesis locally on a gpu

## When to choose
- you need state-of-the-art open-source TTS with strong speaker similarity and low word error rates
- you want voice cloning and emotional expressiveness in a single autoregressive model
- you require high-fidelity 48 kHz output or instruction-controlled speech editing
- you want a pip-installable library with published checkpoints and benchmark-backed quality

## When to avoid
- you need an ultra-light on-device TTS for embedded or CPU-only hardware, since the 2B-parameter model is compute heavy
- you only need a simple rule-based screen-reader voice with minimal resources
- you need a managed cloud TTS API rather than running model inference yourself

## Facets
- artifact type: library
- maturity: active
- function: tts, machine-learning, deep-learning, audio-processing
- domain: speech-processing, artificial-intelligence, deep-learning, machine-learning
- platform: python, cross-platform
- tags: text-to-speech, speech-synthesis, voice-cloning, autoregressive, flow-matching, diffusion, speech-editing, multilingual-tts, 48khz-audio, meanflow, huggingface-checkpoints, gpu

## Member repositories
- studio-dots-ai/dots.tts (main) score 79

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:04:12.851904+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 2026-08-30T05:03:04.121612+00:00, confidence not recorded.
  - readme: https://github.com/studio-dots-ai/dots.tts (fetched 2026-08-28T04:04:12.851904+00:00, sha 814982c44293)
  - registry_pypi: https://pypi.org/pypi/dots.tts/json (fetched 2026-08-29T12:14:15.018478+00:00, sha dcd526552e98)
- Data as of 2026-08-30T08:39:29.467469+00:00.