Graphon vs AgentRun

Agent workflow engines compared

Graphon is a Python engine for defining and executing agent workflows as graphs, while AgentRun is a TypeScript platform for managing agents, revisions, and runs. Graphon centers on graph execution and integrations; AgentRun adds a server and workflow DSL for inspecting and managing runs.

GraphonAgentRun
LanguagePythonTypeScript
LicenseApache-2.0Apache-2.0
Stars86423
Forks4534
Last analyzedNov 9, 2026Nov 2, 2026

Key differences

  • Graphon uses Python and graph-based, queue-driven execution; AgentRun uses TypeScript workflows with typed decisions and agent tasks.
  • Graphon provides workflow state, built-in nodes, and integration protocols for HTTP, files, tools, and human input; AgentRun stores agent revisions and run records behind a server API and interface.
  • Graphon is aimed at developers building or extending agent workflows; AgentRun also targets teams that want to review runs or manage agents through a self-hosted platform.
  • AgentRun supports Docker-based server hosting and durable run history; Graphon describes an extensible engine and integration protocols rather than a server deployment model.
  • Both projects use Apache-2.0. Graphon requires Python 3.12 or later in the 3.12 series, or Python 3.13; AgentRun is written in TypeScript and its published packages use the beta tag.
  • Graphon reports ongoing development with contributors not concentrated around one person; AgentRun is very new, with most contributions from one person and a prerelease as its latest listed release.

Choose Graphon if you…

  • need a Python engine that executes workflows as validated graphs with shared state.
  • want to extend execution with custom engine layers or external command channels.
  • need workflow integration points for HTTP, files, tools, or human input.
Read the Graphon analysis →

Choose AgentRun if you…

  • want to manage agents, revisions, and run records through a server.
  • need workflows with typed decisions, branching, parallel work, or bounded loops.
  • want to inspect which agent revision ran, or use the Pi extension to build and run workflows.
Read the AgentRun analysis →

Compare the numbers side by side, or add more projects

This comparison is generated with AI from the OSRepos analyses of both projects. Always check each project's repository and documentation before choosing.