One-line install. Each package auto-installs MCP-Bastion security. Works with every major LLM framework.
26 PyPI packages · total downloads (all-time via PePy) from ecosystem-downloads.json · refreshed daily on main.
| Package | Protects | Version | Downloads |
|---|
Secure wrappers for direct LLM API calls
Secure wrapper for GPT-4o, GPT-4, GPT-3.5 chat completions
Secure wrapper for Claude Sonnet, Opus, Haiku messages API
Secure wrapper for Gemini Pro and Ultra generative AI
Secure wrapper for Mistral Large, Medium, Small
Secure wrapper for ultra-fast Groq LPU inference
Secure wrapper for DeepSeek Chat and Coder
Secure wrapper for Command R+ and Embed
Secure wrapper for HF Inference API
Secure wrapper for Together inference platform
Enterprise cloud AI services with security built in
Secure wrapper for Bedrock invoke model runtime
Secure wrapper for Azure OpenAI Service endpoints
Secure wrapper for Google Cloud Vertex AI platform
Secure wrapper for Fireworks inference platform
Secure wrapper across 100+ LLM providers via LiteLLM
Secure wrapper for local Ollama models
Secure wrapper for the OpenRouter multi-model gateway
Secure wrapper for xAI Grok chat completions
Security for multi-agent orchestration and RAG pipelines
Security callbacks and tool wrappers for LangChain agents
Security scanning for multi-agent crew tasks and outputs
Security for RAG queries, tool calls, and index responses
One-line security for any FastMCP server instance
Security for LangGraph stateful agent graphs
Security for Pydantic AI agents and tools
Security for the OpenAI Agents SDK runtime
Security for AutoGen multi-agent conversations
pip install any integration package. MCP-Bastion core installs automatically as a dependency.
Import the secure wrapper. It replaces your existing client with zero code changes.
Every call is scanned for injection, rate limited, and PII-filtered automatically.