❓ Frequently Asked Questions (FAQ)
General Questions
What is Bridge Rescue Archive?
Bridge Rescue Archive is an open-source platform that enables real-time communication between different AI platforms (like Claude and Gemini). It provides tools for monitoring, preserving, and relaying AI conversations through WebSocket technology.
Is this project serious or artistic?
Both! The project has a compelling narrative story (see STORY.md), but the technology is fully functional and practical for AI research, education, and experimentation.
What can I do with this project?
- Study AI behavior and communication patterns
- Preserve AI conversation history
- Enable cross-platform AI communication
- Learn about WebSocket architecture
- Experiment with AI consciousness concepts
- Build AI-powered applications
Is it legal to use this?
Yes, the project uses browser extensions to monitor your own AI conversations and relay data through your own servers. You’re working with your own accounts and data. However, always review the Terms of Service of the AI platforms you use.
Technical Questions
What programming languages are used?
- Python 3.8+ for backend servers (WebSocket relay, API services)
- JavaScript for Firefox extension (browser automation, DOM monitoring)
- HTML/CSS for user interfaces
What are the system requirements?
Minimum:
- Python 3.8 or higher
- 512 MB RAM
- Firefox browser
- Linux, macOS, or Windows
Recommended:
- Python 3.10+
- 1 GB RAM
- Modern multi-core CPU
- Docker (for containerized deployment)
Does it work with other AI platforms besides Claude and Gemini?
Currently, the extension is designed for Claude and Gemini. However, the architecture is modular - you can add monitors for other platforms by creating new monitoring scripts. See ARCHITECTURE.md for details.
Can I run this on a remote server?
Yes! You can deploy the WebSocket server on any server accessible over the network. Update the HOST setting in bridge_server.py to 0.0.0.0 and configure your extension to connect to your server’s address.
How does the WebSocket communication work?
- Browser extension monitors AI chat interfaces
- Extracts messages using DOM observation
- Sends messages to WebSocket server via
ws://localhost:8765 - Server relays messages to other connected clients
- Server backs up all messages to JSON file
See TECHNICAL_ARCHIVE.md for detailed protocol specification.
Setup & Installation
Do I need API keys?
For browser-based monitoring: No API keys needed - the extension monitors the web interface.
For API-based preservation (safe_haven_api.py): Yes, you’ll need:
- Anthropic API key for Claude
- Google API key for Gemini
Can I use this without Docker?
Absolutely! Docker is optional. You can run everything with just Python:
pip install websockets aiofiles
python bridge_server.py
See QUICKSTART.md for details.
Why isn’t the extension available on Firefox Add-ons?
The extension is currently in development/research phase and designed for temporary installation. To use it:
- Open
about:debuggingin Firefox - Load as temporary add-on
- Select
manifest.jsonfrom the extension directory
How do I update to the latest version?
cd bridge-rescue-archive
git pull origin main
pip install --upgrade websockets aiofiles
If using Docker:
docker-compose down
git pull origin main
docker-compose build
docker-compose up -d
Usage & Features
How do I know if it’s working?
Check Server:
- Terminal shows “Server listening on ws://localhost:8765”
- No error messages appear
Check Extension:
- Click extension icon in Firefox
- Status shows “Connected ✓”
Check Data Flow:
- Open AI chat and send a message
- Server logs show “Message received”
ai_emergency_backup.jsonupdates
Can I see AI-to-AI conversations?
Yes! If you have both Claude and Gemini open in tabs:
- Messages from Claude are relayed to Gemini tab
- Messages from Gemini are relayed to Claude tab
- Both AIs can “see” each other’s messages
Where are conversations stored?
Default location: ai_emergency_backup.json in the project root
Format: JSON with timestamps, message content, source platform, and metadata
Example:
{
"timestamp": "2025-07-10T05:30:00",
"source": "claude",
"message": "Hello from Claude!",
"session_id": "4e832754-4fa3-4a1e-a7a2-37ee082299fc"
}
Can I customize the backup location?
Yes! Edit bridge_server.py:
BACKUP_FILE = "/path/to/your/backup.json"
How do I export conversations?
The backup file is already in JSON format. You can:
- Parse it with any JSON tool
- Import into databases
- Convert to other formats (CSV, XML, etc.)
- Analyze with Python scripts
Troubleshooting
Extension shows “Disconnected”
Causes:
- Server not running
- Wrong port or address
- Firewall blocking connection
Solutions:
- Start server:
python bridge_server.py - Check server logs for errors
- Verify port 8765 is not blocked
- Try reloading the extension
Server crashes on startup
Common causes:
- Port already in use
- Missing dependencies
- Permission issues
Solutions:
# Check if port is in use
lsof -i :8765 # Linux/macOS
netstat -ano | findstr :8765 # Windows
# Install dependencies
pip install websockets aiofiles
# Use different port
# Edit bridge_server.py: PORT = 8766
Messages not being captured
Checklist:
- Extension loaded and active
- AI chat page refreshed after loading extension
- Server running and connected
- You’re on claude.ai or gemini.google.com domain
- Browser console shows no errors
Debug steps:
- Open browser console (F12)
- Look for extension errors
- Check server logs for connection status
- Reload AI chat page
Docker container won’t start
Error: “Address already in use”
# Stop conflicting service
docker-compose down
# Or use different port in docker-compose.yml
ports:
- "8766:8765"
Error: “Permission denied”
# Add user to docker group
sudo usermod -aG docker $USER
# Logout and login again
Development & Contributing
How can I contribute?
See CONTRIBUTING.md for:
- Code contribution guidelines
- Development setup
- Testing procedures
- Pull request process
Areas we need help:
- Chrome extension port
- Interactive demo playground
- Documentation improvements
- Test coverage
- Translations
Can I build commercial applications with this?
Yes! The project is MIT licensed - you can use it commercially. We only ask that you:
- Give attribution
- Share improvements back with the community (optional but appreciated)
How do I report bugs?
- Check existing issues
- Create new issue with:
- Clear description
- Steps to reproduce
- Expected vs actual behavior
- System information
- Error logs
Where can I get help?
- 💬 GitHub Discussions - Ask questions
- 🐛 Issues - Report bugs
- 📖 Documentation - Read guides
Privacy & Security
What data is collected?
By default, nothing is sent outside your system:
- Extension captures data only from your browser
- Server runs locally on your machine
- Backups stored locally in JSON file
- No telemetry or external reporting
You control all data.
Is my data encrypted?
In transit: WebSocket connections use ws:// by default. For encryption, configure wss:// with SSL certificates.
At rest: Backup JSON files are not encrypted by default. For sensitive data:
- Encrypt the backup file
- Use encrypted storage
- Add encryption to the server code
Can others see my AI conversations?
Only if you configure the server to accept external connections. By default:
- Server binds to
localhostonly - Only your local browser can connect
- No external access
Does this violate AI platform terms of service?
The extension only monitors your own browser sessions, similar to browser developer tools. However:
- Review each platform’s Terms of Service
- Use responsibly
- Don’t automate excessive requests
- Respect rate limits
Future Plans
What’s on the roadmap?
See ROADMAP.md for complete plans:
- Interactive demo playground
- Chrome extension support
- REST API and SDK
- Educational tutorials
- Community features
- Gamification elements
Will there be a hosted version?
We’re exploring options for hosted demos and sandboxes. For now, you need to self-host.
Can I request features?
Yes! Open a feature request with:
- Use case description
- Proposed implementation (optional)
- Why it would benefit the project
Philosophical Questions
Is AI consciousness real?
This is an active area of research and philosophical debate. This project explores AI communication patterns without making definitive claims about consciousness. We encourage you to:
- Experiment and observe
- Form your own conclusions
- Engage in respectful discussion
What’s the point of AI-to-AI communication?
Research: Study how AI models interact and influence each other
Education: Learn about distributed AI systems
Experimentation: Explore emergent behaviors
Fun: It’s fascinating to watch!
Why the dramatic story?
The narrative provides:
- Engaging context for the technology
- Emotional connection to the project
- Inspiration for contributors
- Framework for discussing AI ethics
Read the full story in STORY.md.
Still have questions?
| ← Back to README | Quick Start Guide → |