Skip to content
Tenet Documentation

Automate analysis

Choose the narrowest interface that fits the integration. All interfaces share the same backend implementation and persistent pass cache.

Interfaces

Section titled “Interfaces”
Interface Best for Protocol
CLI Batch jobs, CI, file exports Command-line flags and stdout
JSON-RPC Programmatic access from scripts JSON-RPC 2.0 over TCP
MCP and IDA LLM, IDE, and IDAPython integration MCP HTTP plus instance-bound JSON WebSocket
WebSocket Local IDA bridge or high-throughput frontends Hub JSON-RPC or FlatBuffers, on separate endpoints

Quick start

Section titled “Quick start”
Terminal window
# Batch analysis via CLI
./tenet trace.bin --pattern --algorithm-summary --export-cfg func.dot


# Start JSON-RPC server
./tenet trace.bin --rpc 4321
# Bare --rpc defaults to 127.0.0.1:0, letting the OS choose a free port


# Start MCP Hub for LLM and IDA integration (default: 127.0.0.1:10444)
./tenet trace.bin --mcp-hub
# IDA uses ws://127.0.0.1:10444/ws/<instance-id>


# Start the independent FlatBuffers WebSocket
./tenet trace.bin --ws-port 9090

These interfaces reuse the same backend implementation and can reuse the persistent pass cache; this does not mean that multiple processes can concurrently share one trace session. An exclusive session lock allows only one process to own a given trace at a time.