Integrations

Connect your AI in one line. No SDK required.

LLM Proxy (OpenAI / LangChain / CrewAI)

Point any OpenAI-compatible client at SAILLENT. All prompts & responses are inspected.

client = OpenAI(base_url="http://localhost:3000/v1")

MCP Gateway (Claude / Cursor / Tools)

Route tool calls through SAILLENT. Every MCP action is authorized by the policy engine.

{ "mcpServers": { "saillent": { "url": "http://localhost:3000/mcp" } } }

Python SDK (Custom Agents)

For bespoke agents that call tools directly.

pip install saillent
from saillent import Firewall
fw = Firewall("http://localhost:3000")
if fw.check(agent="bot", tool="db", action="Read", payload=q).allowed: run(q)

Universal Coverage

Zero-trust enforcement across every major model, framework, and IDE.

All Systems Monitored

OpenAI

GPT-5 / o3

Anthropic

Claude 3.5

Frameworks

LangChain / Llama

IDEs

Cursor / Cline

Multi-Agent

CrewAI / AutoGen

Local Models

Ollama / LM Studio

Vercel AI SDK

React / Next.js

Custom Tools

REST / GraphQL