# Authentication Guide
Some semiconductor B2B platforms require authentication to access their data. This MCP server supports multiple authentication methods.
## Supported Sites
### DesignReuse
- Requires username and password
- Session cookies are automatically managed
- Credentials stored locally in `~/.semiconductor-mcp/auth.json`
### AnySilicon
- Currently public access (no auth required)
- API key support planned for premium features
## Configuration Methods
### 1. Environment Variables
Set credentials via environment variables:
```bash
export DESIGNREUSE_USERNAME="your_username"
export DESIGNREUSE_PASSWORD="your_password"
```
### 2. Configuration File
Create `~/.semiconductor-mcp/auth.json`:
```json
{
"designreuse": {
"username": "your_username",
"password": "your_password"
}
}
```
### 3. Interactive Setup
Use the `setup_auth` tool:
```
Tool: setup_auth
Arguments: {"site": "designreuse"}
```
### 4. Pass Credentials in Tool Calls
Some tools accept optional auth parameters:
```json
{
"tool": "find_ip_vendors",
"arguments": {
"category": "PHY",
"auth": {
"username": "user",
"password": "pass"
}
}
}
```
## Security Considerations
1. **Local Storage**: Credentials are stored locally in your home directory
2. **Encryption**: Consider using OS keychain integration (planned feature)
3. **Session Management**: Sessions are cached to minimize login requests
4. **No Cloud Storage**: Credentials never leave your machine
## Handling Authentication Failures
If authentication fails:
1. Check credentials are correct
2. Verify account is active on the platform
3. Some sites may require CAPTCHA (use headless: false for manual intervention)
4. Check for IP rate limiting
## Best Practices
1. Use dedicated API accounts when available
2. Rotate credentials regularly
3. Monitor usage to avoid rate limits
4. Use caching to minimize requests
## Troubleshooting
### "Authentication required" error
- Run `setup_auth` tool to configure credentials
- Check `~/.semiconductor-mcp/auth.json` exists and is valid
### "Login failed" error
- Verify credentials are correct
- Try logging in manually to check for CAPTCHA or account issues
- Set `headless: false` in scraper options for debugging
### Session expires frequently
- Normal behavior for security
- Sessions auto-renew on next request
- Consider implementing refresh token logic
