Trango ComputeContextIQPreview
Free tool

See how your
agents connect

Paste any GitHub repository URL to instantly visualize how AI agents, tools, and workflows connect. Supports LangGraph, CrewAI, AutoGen, Google ADK, and OpenAI Agents SDK.

Frameworks detected

LangGraphCrewAIGoogle ADKAutoGenOpenAI Agents SDKSwarmStateGraphGroupChatManagerRoundRobinGroupChatSequentialAgentParallelAgentRunner.run

What it does

Everything you need to understand
any agent codebase

Scan any GitHub repo

Paste a public GitHub URL and the tool fetches the Python source, detects the framework, and builds the graph in seconds. No API key or local setup required.

5 frameworks supported

LangGraph StateGraph, CrewAI Crew, Google ADK LlmAgent, AutoGen GroupChatManager, and OpenAI Agents SDK / Swarm are all detected automatically from imports and class usage.

Node inspector

Click any node to see the agent's LLM model, bound tools, system prompt excerpt, and exact source file and line number — without leaving the browser.

Layered & compact layouts

Switch between layered view (routers on their own row) and compact view (routers and agents share a row) to fit any workflow size on screen.

Export to PNG & SVG

Download the graph as a high-resolution PNG or vector SVG for documentation and presentations. Pro feature. Pro feature.

Export graph as JSON

Download the full GraphSchema as JSON — nodes, edges, metadata, framework, and commit SHA — free for every user.

Node types in the graph

Agent

LLM-powered agent with model and tools

Tool

@tool decorated function or ToolNode

Router

Conditional edge router or orchestrator

State / Memory

MemorySaver, SqliteSaver, state schema

Entry (START)

Graph entry point

Exit (END)

Graph exit point

Guardrail

Input/output guard or safety filter

Parallel

Agents running concurrently

Who uses it

AI engineers onboarding to a new codebase

Understand a LangGraph or CrewAI repo in minutes instead of hours. See every agent, tool, and handoff without reading hundreds of lines of Python.

Team leads reviewing agent architecture

Spot dead ends, missing tools, or unintended routing in a pull request before it ships. The graph makes structural issues immediately visible.

Researchers comparing multi-agent frameworks

Scan reference implementations of LangGraph, AutoGen, Google ADK, and OpenAI Agents side by side to understand how each framework structures agent communication.

Try Agent Workflow Visualizer free →

Open the toolSee all ContextIQ tools