AISBF Gets Major Update: TOR Hidden Services, MCP Server, and Kiro Support

AISBF (AI Service Broker Framework) just received a massive update with several game-changing features. Here's what's new:

🚀 Key New Features

TOR Hidden Service Support (v0.5.0)

You can now expose AISBF over the TOR network as a hidden service, providing anonymous access to your AI proxy server. Perfect for privacy-conscious deployments.

MCP Server Endpoint

Full Model Context Protocol support for remote agent configuration and model access via SSE and HTTP streaming. This opens up new possibilities for AI agent integrations.

Kiro Provider Support

Added support for Kiro (Amazon Q Developer / AWS CodeWhisperer) as a new AI provider. AISBF now supports:

Google (google-genai)
OpenAI and openai-compatible endpoints
Anthropic
Ollama
Kiro (Amazon Q Developer)

Web Dashboard Enhancements

OpenRouter-style extended fields in the Model class
Comprehensive dashboard documentation
New "Create Persistent" button for TOR settings
Proxy-awareness and configurable error cooldown features

Python 3.13 Compatibility

Fixed multiple compatibility issues with Python 3.13, including Jinja2 template caching and session references.

PyPI Package Fixes

Fixed distribution to include mcp.py, tor.py, and kiro modules in the package.

Why AISBF?

AISBF is a modular proxy server that manages multiple AI provider integrations with a unified API. It offers:

Intelligent Routing: Weighted load balancing across providers
Autoselect Models: AI-powered model selection based on content
Streaming Support: Full streaming from all providers
Token Rate Limiting: Per-model TPM/TPH/TPD tracking
Context Management: Automatic condensation when approaching limits
SSL/TLS: Built-in HTTPS with Let's Encrypt integration

Get Started

pip install aisbf

Then run aisbf and access the dashboard at http://localhost:17765/dashboard (default credentials: admin/admin)

Nexlab