https://github.com/stanford-mast/blast

Browser-LLM Auto-Scaling Technology
https://github.com/stanford-mast/blast

ai-agents browser-automation llm-inference python

Last synced: 6 months ago
JSON representation

Browser-LLM Auto-Scaling Technology

• Host: GitHub
• URL: https://github.com/stanford-mast/blast
• Owner: stanford-mast
• License: mit
• Created: 2025-04-05T00:03:34.000Z (8 months ago)
• Default Branch: main
• Last Pushed: 2025-05-08T18:59:10.000Z (7 months ago)
• Last Synced: 2025-05-08T19:46:08.777Z (7 months ago)
• Topics: ai-agents, browser-automation, llm-inference, python
• Language: Python
• Homepage: http://blastproject.org/
• Size: 3.23 MB
• Stars: 487
• Watchers: 7
• Forks: 17
• Open Issues: 4
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

• awesome-llm-os - BLAST: Browser-LLM Auto-Scaling Technology
• awesome-LangGraph - stanford-mast/blast - performance serving engine for browser-augmented LLM applications with auto-scaling and OpenAI-compatible API |  | (🌐 Web Automation & Scraping / 🟩 Development Tools 🛠️)

README

          


  



A high-performance serving engine for web browsing AI.




[![Website](https://img.shields.io/badge/blastproject.org-FFE067)](https://blastproject.org)

[![Documentation](https://img.shields.io/badge/Docs-FFE067)](https://docs.blastproject.org)

[![Discord](https://img.shields.io/badge/Discord-FFE067)](https://discord.gg/NqrkJwYYh4)

[![Twitter Follow](https://img.shields.io/twitter/follow/realcalebwin?style=social)](https://x.com/realcalebwin)





  



## ❓ Use Cases

1. **I want to add web browsing AI to my app...** BLAST serves web browsing AI with an OpenAI-compatible API and concurrency and streaming baked in.

2. **I need to automate workflows...** BLAST will automatically cache and parallelize to keep costs down and enable interactive-level latencies.

3. **Just want to use this locally...** BLAST makes sure you stay under budget and not hog your computer's memory.

## 🚀 Quick Start

```bash

pip install blastai && blastai serve

```

```python

from openai import OpenAI

client = OpenAI(

    api_key="not-needed",

    base_url="http://127.0.0.1:8000"

)

# Stream real-time browser actions

stream = client.responses.create(

    model="not-needed",

    input="Compare fried chicken reviews for top 10 fast food restaurants",

    stream=True

)

for event in stream:

    if event.type == "response.output_text.delta":

        print(event.delta if " " in event.delta else "", end="", flush=True)

```

## ✨ Features

- 🔄 **OpenAI-Compatible API** Drop-in replacement for OpenAI's API

- 🚄 **High Performance** Automatic parallelism and prefix caching

- 📡 **Streaming** Stream browser-augmented LLM output to users

- 📊 **Concurrency** Out-of-the-box support many users with efficient resource management

## 📚 Documentation

Visit [documentation](https://docs.blastproject.org) to learn more.

## 🤝 Contributing

Awesome! See our [Contributing Guide](https://docs.blastproject.org/development/contributing) for details.

## 📄 MIT License

As it should be!