Model reference · open weights
Audio8-TTS is an open-weight audio or speech model from Audio8, listed in the AxForge catalogue. AxForge can bring it up on EU-owned hardware for you on request — with the licence handled where one is required.
About
A 0.6B-parameter multilingual text-to-speech model with zero-shot voice cloning. [](https://github.com/Audio8-AI/Audio8TTS) [](https://audio8-ai.github.io/Audio8TTS/) [](https://huggingface.co/Audio8/Audio8-TTS-Preview-0.6B-ONNX-INT4) [](https://github.com/Audio8-AI/Audio8TTS/blob/main/LICENSE) Audio8 TTS Preview supports multilingual speech generation and zero-shot voice cloning. This repository contains the complete checkpoint, its 44.1 kHz neural audio codec, tokenizer, processor, and Hugging Face remote code. Preview status: Language coverage is intentionally limited in this release. For the best results, use one of the 11 recommended languages below. Broader multilingual coverage and Chinese dialect support are planned for future releases. Supported Languages Model Details Audio8 TTS uses a DualAR architecture inspired by Fish Audio S2 Pro. The slow AR transformer predicts one semantic token for each audio frame. The fast AR transformer predicts the frame's codec codebooks, conditioned on the slow hidden state and preceding codebooks. The bundled codec handles both reference-audio encoding and waveform decoding, so no additional codec checkpoint is required. Installation Python 3.10 or newer and a CUDA-capable GPU are recommended. Usage The model uses custom Transformers code. Review the files in this repository, then load it with trustremotecode=True. Zero-shot voice cloning The reference transcript must match the spoken content in the reference audio. Generation without a reference Omit referenceaudio and referencetext when a cloned voice is not needed: For command-line inference, batching, and supervised fine-tuning, see the Audio8 TTS repository. Deployment Options CPU deployment: ONNX INT4 Audio8-TTS-Preview-0.6B-ONNX-INT4 packages Audio8 TTS for low-resource CPU inference with ONNX Runtime. Slow and Fast AR weights use weight-only INT4, while activations, KV caches, and the neural audio codec use FP16. Normal synthesis loads only the Slow AR, Fast AR, and codec decoder sessions. Voice registration releases those sessions before loading the optional codec encoder, keeping peak memory controlled. Get the ONNX INT4 model and follow the CPU ONNX Runtime g
Summarised from the published model card. Read the full card on the HuggingFace links below.
Specifications
| Maker | Audio8 |
|---|---|
| Type | Audio & music |
| Parameters (lead) | 601M |
| Variants | 2 |
| Runs with | transformers |
| Released | 2026-07-28 |
| Popularity | 21k downloads / month |
| Likes | 381 |
| Licence | Commercial licence needed |
How it works
Variants
Open weights ship in several sizes and precisions. One page, all the variants — pick the one that fits your GPU. VRAM figures are estimates from model size.
Using it via the API
Once AxForge deploys audio8-tts for you, it answers on the OpenAI-compatible API — the same base URL and keys as every other model. (audio8-tts below is illustrative; you get the exact model name on deployment.)
$ curl -sS https://api.axforge.ai/v1/audio/transcriptions \ -H "Authorization: Bearer $AXFORGE_API_KEY" \ -F model="audio8-tts" -F file=@audio.mp3
Details
Languages
Tags
Licence
The weights are open but apache-2.0 needs a commercial agreement for business use. AxForge can arrange that licence and host the model for you — you pay AxForge, we settle with the model’s maker. Ask us for a quote. Read the licence ↗