Skip to the content
Know your tech before you buy it. Makers: get listed from $49/mo

Fish Speech

github.com

  • Free plan
  • Linux
  • Web
  • API
  • Self-hosted
7.5 Club score Visit github.com

What Fish Speech is

Fish Speech is an open-source multilingual text-to-speech system for generating expressive speech. Its S2 model accepts natural-language tags to control emotion and prosody. Voice cloning works from short reference samples, typically 10–30 seconds, without additional fine-tuning; multiple reference speakers and speaker ID tokens can also be used to generate dialogue in one pass. The repository describes S2 Pro as supporting more than 80 languages without phonemes or language-specific preprocessing.

Fish Speech can be run from the command line, through a WebUI or server, or with Docker. API access is available, and documented setups include CPU installation and GPU deployment. The installation guide lists Linux or WSL and 24 GB of GPU memory for inference. The research license allows research and non-commercial use free of charge; commercial use requires a separate written license. Plans include a free tier with 8,000 monthly credits and paid tiers beginning at $11 a month.

Who it suits

Fish Speech suits developers and other users building multilingual speech generation, voice cloning, or multi-speaker dialogue workflows. Commercial users need a separate written license, and inference requires the listed Linux or WSL environment and GPU memory.

Good to know, and what to watch

  • Inline tags control emotion and prosody
  • Clones voices from short reference samples
  • Supports multi-speaker dialogue generation
  • S2 Pro supports more than 80 languages
  • Offers command-line, WebUI, server, and Docker setups
  • Commercial use requires a separate written license
  • Inference guide lists 24 GB GPU memory
  • Inference system requirements include Linux or WSL

Enterprise

—

Free Tier

Free

Plus

$11/mo

Pro

$75/mo

Open source

Getting started

  1. Open the Fish Speech GitHub repository
  2. Choose command-line, WebUI, server, or Docker installation
  3. Use Linux or WSL for the documented system environment
  4. Install for CPU use or deploy with a GPU
  5. Run the system locally or connect through its API

Limits to check

The Free Tier provides 8,000 credits monthly, up to seven minutes of generation, 500 characters per generation, three public voice slots, and standard speed. Commercial rights are excluded from the research license and require a separate written license.

The verdict

Fish Speech provides several ways to deploy expressive, multilingual speech generation, including voice cloning and multi-speaker dialogue. Its non-commercial license terms and listed inference requirements matter before adopting it.

Questions people ask

Is Fish Speech free?

The software is free and open source for research and non-commercial use. A hosted Free Tier costs nothing and includes 8,000 monthly credits.

Can Fish Speech be used commercially?

No commercial rights are granted by the model license. Commercial use requires a separate written license from Fish Audio.

Which platforms are supported?

Listed platforms are API, Linux, self-hosted, and web. Installation documentation covers command-line tools, WebUI, server, and Docker.

Does it support voice cloning?

Yes. S2 can clone a voice from a short 10–30-second reference sample without additional fine-tuning.

What does the Plus plan cost?

Plus costs $11 a month, billed as 132 USD annually, with 250,000 monthly credits.

What the maker publishes

Commercial limits
The model license does not grant commercial rights, which require a separate written license from Fish Audio. source
Emotion and prosody control
S2 supports fine-grained inline control of prosody and emotion through natural-language tags. source
Hardware requirement
The installation guide lists 24 GB of GPU memory for inference and Linux or WSL as system requirements. source
Integrations
The enterprise page lists support for real-time pipelines, WebRTC infrastructure, workflow automation, Vapi, Twilio, and Retell. source
Languages
The repository describes S2 Pro as supporting more than 80 languages without phonemes or language-specific preprocessing. source
License
The Fish Audio Research License permits research and non-commercial use free of charge, while commercial use requires a separate written license. source
Multiple speakers
S2 can use multi-speaker reference audio and speaker ID tokens to generate dialogue with multiple speakers in one generation. source
Purpose
Fish Speech is an open source multilingual text-to-speech system that generates expressive speech. source
Security and compliance
Fish Audio says its SOC 2 Type II audit is underway and offers zero data retention, HIPAA-aligned configurations, and self-hosted deployments for enterprise workloads. source
Streaming
The documentation describes SGLang serving optimizations including continuous batching, paged KV cache, CUDA graphs, and prefix caching. source
Support
Enterprise support includes 24/7 production support, a technical account manager, and a 99% uptime SLA. source
Voice cloning
S2 supports voice cloning from short reference samples, typically 10–30 seconds, without additional fine-tuning. source
Ways to run
The installation guide documents command-line, WebUI, server, and Docker setups, including CPU installation and GPU deployment. source

Checked against the maker's pages on 5 Oct 2026.