Trust Enforcement for Every AI Framework
Native integrations that drop into your existing stack. Verify agent identity, enforce security policies, and emit audit events — in two lines of code.
Framework Integrations
First-class libraries purpose-built for each framework.
LangChain & LangGraph
Composable LCEL guard that slots into any chain or graph. Zero config — just pipe and go.
pip install langchain-capisciofrom langchain_capiscio import CapiscioGuard
from langchain_openai import ChatOpenAI
safe_chain = CapiscioGuard() | ChatOpenAI(model="gpt-4o-mini")
result = safe_chain.invoke("Summarise quarterly earnings")
CrewAI
Protect multi-agent crews with per-agent trust badges. Verify every delegation before it executes.
pip install capiscio-sdkfrom capiscio_sdk.simple_guard import SimpleGuard guard = SimpleGuard() # auto-discovers agent-card.json guard.verify(request) # identity + integrity + freshness
MCP (Model Context Protocol)
Validate tool-use requests between AI models and external tools. Guard every MCP call with cryptographic proof.
pip install capiscio-mcpfrom capiscio_mcp import McpGuard guard = McpGuard() # Validates MCP tool calls against trust policies guard.protect(mcp_server)
SDKs & CLIs
Core libraries for building custom integrations.
Python SDK
Full-featured SDK with SimpleGuard, FastAPI middleware, and async support.
pip install capiscio-sdkGo Library
High-performance Go library with HTTP middleware and gRPC interceptors.
go get github.com/capiscio/capiscio-core/v2Don't See Your Framework?
The Python SDK and Go library work with any HTTP framework. Or request a native integration — we ship fast.