AISBF
a self-hostable OpenAI-compatible AI proxy/router
Last verified:
What is AISBF?
AISBF (AI Service Broker Framework) is a unified AI gateway and proxy server that provides intelligent routing, load balancing, and failover across multiple AI providers. It offers a single unified API for accessing Google, OpenAI, Anthropic, Ollama, and other AI services, allowing developers to switch providers without changing their code. The tool is designed for developers, teams, and organizations building resilient AI applications who want to reduce costs, avoid vendor lock-in, and maintain privacy control over their AI infrastructure.
Key features include AI-powered model selection based on content, automatic context condensation using four methods (hierarchical, conversational, semantic, and algorithmic), built-in rate limiting with token tracking (TPM/TPH/TPD), response caching with intelligent invalidation, and full streaming support via WebSocket and Server-Sent Events. AISBF includes multi-user support with individual API keys, quotas, and usage tracking, making it perfect for teams and projects. It also offers native TOR hidden service support for anonymous access and complete privacy.
AISBF is ideal for developers tired of API limits, teams managing multiple AI providers, organizations needing privacy-focused AI infrastructure, and anyone wanting to mix local LLMs (via Ollama) with cloud providers. The platform claims to save up to 70% on API costs through intelligent routing, caching, and context management. It's open source and self-hosted, giving users full control over their infrastructure.
AISBF pricing
Pricing model: Freemium
AISBF is free and open source. Self-host for free by viewing source code on GitLab. Both AISBF and LiteLLM are excellent OSS solutions. There is no paid tier - the tool is completely free as open source software that you self-host on your own infrastructure.
AISBF pros
- Free and open source (OSS)
- Intelligent routing with AI-powered model selection
- Built-in dashboard included
- Multi-user support with virtual keys
- Native rate limiting without extra config
- Built-in response caching with intelligent invalidation
- TOR hidden service for anonymous access
- Automatic context condensation with 4 methods
- Seamless local Ollama + cloud provider mixing
- Auto-failover when providers fail or hit limits
- Detailed analytics per user, model, and provider
- WebSocket and SSE streaming support
- OpenAPI compatible - drop-in OpenAI replacement
- Self-hosted with full infrastructure control
- Zero vendor lock-in - switch providers anytime
- Save up to 70% on API costs
- Multi-provider support in unified interface
- Token tracking (TPM/TPH/TPD)
- Simple REST API with Python SDK
- Handle thousands of requests with load balancing
AISBF cons
- Self-hosting requires technical setup knowledge
- Active development may have bugs
- Smaller community than established alternatives
- TOR access adds complexity for some users
- Requires own infrastructure/deployment
- Less documentation than major commercial APIs
- Local model quality depends on user's hardware
- Auto-selection may not match specific preferences
Frequently asked questions about AISBF
What is AISBF?
AISBF (AI Service Broker Framework) is a unified AI gateway providing intelligent routing, load balancing, and failover across multiple AI providers. It offers a single unified API for Google, OpenAI, Anthropic, Ollama, and more, with the motto 'AI Should Be Free'.
Is AISBF free?
Yes, AISBF is free and open source. You can self-host it for free by viewing the source code on GitLab. There are no paid plans - it's completely free OSS.
How does intelligent routing work?
AISBF uses weighted load balancing, automatic failover, and AI-powered model selection based on your content. It autoselects the best model for your specific prompt and context.
What providers does AISBF support?
AISBF supports Google, OpenAI, Anthropic, Ollama, and other compatible services through a unified interface. You can switch providers without changing your code.
How does context condensation work?
AISBF offers automatic context condensation using four methods: hierarchical, conversational, semantic, and algorithmic. This helps reduce token usage and costs.
Can I use AISBF anonymously?
Yes, AISBF has native TOR hidden service support. You can access it via TOR (aisbfity4ud6nsht53tsh2iauaur2e4dah2gplcprnikyjpkg72vfjad.onion) or self-host as a hidden service for complete privacy.
How does AISBF compare to LiteLLM?
Both are free OSS solutions. AISBF has built-in caching (LiteLLM requires config), native TOR support (LiteLLM doesn't have it), autoselect model choice (LiteLLM doesn't), context condensation with 4 methods (LiteLLM doesn't), and auto-failover for Ollama + cloud mixing (LiteLLM needs manual config). Both have dashboards and multi-user support.
Can I use local LLMs with AISBF?
Yes, AISBF has built-in Ollama integration for local LLMs. It can automatically mix local and cloud providers with auto-failover, switching seamlessly between them.
How do I get started with AISBF?
Install via pip install aisbf. The tool has a simple REST API, Python SDK, and extensive documentation to get you started in minutes. You can also view source code on GitLab for self-hosting.
How much can I save with AISBF?
AISBF claims to save up to 70% on API costs through intelligent routing, caching, and context management. Detailed analytics help you track usage per user, model, and provider to optimize your budget.