What changed
The Fish Audio plugin for the Hermes Agent has been released, providing Hermes agents with advanced speech and listening capabilities. This plugin enables expressive voices in over 80 languages, voice design features, and live dictation. It is designed to work across the Hermes CLI, all supported messaging applications, and the Hermes Desktop app. The installation process involves using the Hermes CLI command hermes plugins install fishaudio/fish-audio-hermes --enable --yes-deps. Users are required to obtain a Fish Audio API key, with new accounts eligible for the free s2.1-pro-free model until November 30, 2026. The --yes-deps flag ensures that the plugin's Python dependencies are installed automatically. The plugin supports various features, including expressive speech with emotion and delivery cues, streaming voice output, a library of community voices, and the ability to create custom voices through cloning or descriptive input. For speech-to-text, it utilizes transcribe-1-pro for voice notes and offers realtime dictation with word-by-word captions for live input. The latest release is version v1.14.1.
Why it matters for builders
AI builders can now integrate sophisticated voice interaction into their Hermes agents without needing to develop these capabilities from scratch. The plugin's extensive language support and voice design tools empower developers to create highly personalized and engaging AI personas. The real-time dictation feature also opens up possibilities for interactive command and control interfaces or live transcription services within agent workflows.
Practical impact
Developers can install the plugin using hermes plugins install fishaudio/fish-audio-hermes --enable --yes-deps and then authenticate with their Fish Audio API key via hermes fish login. After setup, users can interact with the agent's voice features through commands like /fish setup to initiate configuration, /fish preview to test voices, or by directly asking the agent to perform voice-related tasks. For scripting, commands such as hermes fish speak and hermes fish transcribe are available for programmatic use. The plugin also offers tools for managing voices, checking account balance, and configuring speech settings.
Caveats and source limits
The source material indicates that new accounts receive free access to the s2.1-pro-free model until November 30, 2026, after which API credit will be required. Some features, such as voice previews and voice designs, are billed by Fish Audio. There is a noted compatibility issue with older Hermes versions (0.23.0) where the agent might fall back to an older catalog version of the plugin, requiring manual removal and reinstallation of the latest release. The source does not provide independent benchmark results for latency or accuracy, relying on Fish Audio's internal measurements for streaming voice. Pricing details beyond the initial free tier are not specified.
Sources
Claim check: 7/7 supported claims - 7 evidence links - 100% avg confidence
- Fish Audio has released an official plugin for the Hermes Agent.supported - github.com
- The plugin enables expressive voices in over 80 languages for Hermes agents.supported - github.com
- The plugin supports voice design and live dictation.supported - github.com
- The plugin integrates with Hermes CLI, messaging apps, and the Desktop app.supported - github.com
- New accounts receive free access to the `s2.1-pro-free` model until November 30, 2026.supported - github.com
- Voice previews and voice designs are billed by Fish Audio.supported - github.com
- The latest release version is v1.14.1.supported - github.com
Caveats
- Single-source caution: verify critical details at the linked source.
Radar score 87/100 - how it was calculated
- Reliability 82: GitHub metadata supports source trust
- Freshness 92: Fresh GitHub activity
- Novelty 77: Fresh GitHub release
- Technical 85: Repository technical metadata
- Developer 96: Developer tooling signals
- Ecosystem 72: Fresh GitHub release
- Confidence 100: Claims have reliable evidence
Discussion
Loading comments...