Our take
Quietno commits in the last 90 days- Activity
- 32
- Community
- 36
- Issues
- 21
- Pull requests
- 21
+1 stars in the last 30 days
BrowserAI is a broad browser-inference toolkit with useful text and audio capabilities, but its quiet recent activity and reliance on one main contributor make it worth evaluating carefully before adopting it as a core dependency.
Good fit if
- You need local browser inference for supported language, speech, or audio models and can target WebGPU-capable browsers.
- Your team is comfortable integrating a TypeScript library and testing its behavior on the devices and browsers you support.
- MIT licensing fits your project, and you can tolerate a project with no commits or issue and pull-request activity in the recent quarter.
Look elsewhere if
- You need server-side inference, broad compatibility with browsers that lack WebGPU, or a deployment that does not depend on users’ hardware.
- Your adoption requires active recent maintenance or distributed contributor ownership. Most contributions come from one person, and no issues or pull requests were closed or merged in the recent quarter.
- You need a project with tests already documented in its repository metadata, which reports none.
All health signals
| Last commit | 2026-07-21 (3 months ago) |
|---|---|
| Commits, last 90 days | 0 |
| Releases, last 12 months | 1 (latest v2.2.0, 2026-04-18) |
| Contributors | 7 (top contributor: 77% of commits) |
| Issues closed, last 90 days | 0 |
| Pull requests merged, last 90 days | 0 |
| Project age | 21 months |
Checked on 2026-10-13 with the GitHub API.
Project history
Daily GitHub snapshots since Oct 7, 2026. History covers stars and forks; health signals are not tracked historically.
Stars
1,451Forks
137Change since Nov 8, 2026: 0 stars·0 forks
Overview
BrowserAI provides a JavaScript API for loading and running supported AI models in a browser. It covers text generation, speech recognition, text-to-speech, embeddings, and audio source separation, with inference handled on the user’s device rather than by a hosted model server.
It is aimed at web developers who want to add local inference to an application and can rely on compatible browsers and user hardware. The library offers several inference engines and model options, but browser capability and the cost of downloading and running a model remain important constraints.
Key Features
- Runs supported models locally in the browser, with WebGPU acceleration where available.
- Supports multiple inference engines, including MLC, Transformers, Flare for GGUF models, and Demucs for audio separation.
- Provides a shared API for loading models and generating text, with options such as system prompts and JSON-schema output.
- Includes speech recognition through Whisper models and text-to-speech through Kokoro.
- Can run model work in Web Workers to avoid blocking the main UI thread.
- Offers a Demucs engine for separating audio into vocal, drum, bass, and other stems.
- Includes database support for storing conversations and embeddings.
Use Cases
- A web-app developer can add local chat or text generation when keeping prompts on the user’s device is a priority.
- A team building browser-based voice features can combine speech recognition and text-to-speech without routing audio through its own inference server.
- Researchers and hobbyists can experiment with small language models in a browser, provided their browser and hardware support the selected model.
- Audio-tool developers can use the Demucs engine to offer browser-based stem separation for uploaded audio.
What you need
Detected in the repository
- Node.js (from package.json)
- Automated checks on GitHub Actions
License in plain words
MITpermissive
- Commercial use: yes
- Modify and redistribute: yes
- You must keep: the copyright and license notice
- Share your changes: no
A summary, not legal advice: the LICENSE file is what applies.
Getting Started
Install the package:
npm install @browserai/browserai
See the README and documentation for model loading and API examples.
Considerations
The README requires a modern browser with WebGPU support, citing Chrome 113+, Edge 113+, or an equivalent browser. Some models also require hardware support for 16-bit floating-point operations. Model downloads and inference use local device resources, so performance and feasibility can vary by device; the README does not provide benchmark guarantees.
Repository health is marked Quiet: there have been no commits in the recent quarter, despite a release during the last year. Recent issue and pull-request handling is also inactive, and contributions are concentrated in one of the seven listed contributors. The repository metadata reports CI but no tests. The README’s roadmap describes planned work such as enhanced RAG, observability, and enterprise features, so do not assume those roadmap items are already available.
Found this useful?
Share it with someone who would like BrowserAI.