Fish Audio
Fish Audio is an AI text-to-speech and voice cloning tool. In 2026 it added the S2-Pro model with bracket-syntax emotion control and a free plan.
About Fish Audio
Fish Audio is a text-to-speech and voice cloning platform that generates natural-sounding speech from a short voice sample, in the Audio & Music category of AI tools.
What is Fish Audio (2026 update)
Fish Audio lets you type text and hear it read back in a chosen voice, or clone a new voice from as little as 10 to 30 seconds of reference audio. It supports over 50 languages with cross-lingual and code-switching speech, meaning one cloned voice can read a script that mixes languages in the same sentence.
What is new in 2026
On March 9, 2026, Fish Audio open-sourced S2-Pro, built on Qwen3-4B, and released S2 as its next-generation TTS model. The headline change is emotion control written directly into the text using a bracket syntax, for example [whisper], [laugh], or [emphasis], with more than 15,000 tags covering emotion, tone, volume, and rhythm. Fish Audio also launched Story Studio, an editor for longer audio projects where you can add pauses, manage multiple speakers, and regenerate individual clips instead of the whole file.
Key features
- Instant voice cloning from a 10 to 30 second sample
- S2 and S2-Pro models with bracket-syntax emotion and paralinguistic control
- Multi-speaker support in a single project
- 80+ languages supported by the S2 model, with cross-lingual code-switching
- Story Studio for long-form narration with pause and speaker management
- API access priced per character for developers building TTS into their own apps
- Community voice library for using shared voices instead of cloning your own
Pricing in 2026
Fish Audio runs a free plan plus several paid subscription tiers, and a separate per-character API rate for developers. Some sources report an early-2026 promotional discount on the Plus and Pro plans; check the official site for the current live price before subscribing.
| Plan | Price | What you get |
|---|---|---|
| Free | Free | 8,000 credits per month (about 7 minutes of audio), community voice library, basic voice cloning, personal non-commercial use only |
| Plus | $15/month ($66/year billed annually) | 250,000 monthly credits, API access |
| Pro | $100/month ($450/year billed annually) | 2,000,000 monthly credits, 3 team seats, more voice clone slots |
| Max | Custom (reported around $8,988/year) | Highest usage tier for teams and heavy API use, check the official site |
The API is billed separately at roughly $15 per million UTF-8 bytes (about $15 per million English characters), which is the rate to check if you plan to call Fish Audio programmatically rather than use the web app.
Who should use it
Fish Audio fits creators, podcasters, and developers who need natural, emotionally expressive narration without hiring voice actors, especially anyone who wants fine control over how a line is delivered through bracket tags rather than picking from a fixed list of ‘moods.’ The free tier is enough to test voice cloning quality before committing to a paid plan, but any monetized video, podcast, or product needs a paid subscription since the free plan is personal-use only.
Limitations and alternatives
The free plan’s credit allowance is small, around 7 minutes of generated audio per month, so regular creators will hit the ceiling fast. Commercial use is blocked on the free tier, and exact current pricing has shifted with promotions, so confirm the live price on the official site before budgeting. Three real alternatives worth comparing are ElevenLabs, which is positioned as a pricier competitor that Fish Audio’s API undercuts on cost, Smallest.ai, and standard cloud TTS options from Descript for creators who want editing and narration in one tool.
Frequently asked questions
Is Fish Audio free to use?
Yes, there is a free plan with 8,000 credits per month, roughly 7 minutes of generated audio, but it is limited to personal, non-commercial use.
What is new in Fish Audio's 2026 models?
Fish Audio released S2 and open-sourced S2-Pro in March 2026, adding bracket-syntax emotion and paralinguistic control (like [whisper] or [laugh]) across more than 15,000 tags.
How much does the Fish Audio API cost?
The API is billed at roughly $15 per million UTF-8 bytes, about $15 per million English characters, separate from the subscription plans.
Can Fish Audio clone a voice from a short sample?
Yes, it can clone a voice from as little as 10 to 30 seconds of reference audio and generate speech in that voice across 50+ languages.
Information last verified: September 2026.
Related Tools
