The declarative standard for multi-agent systems

Building one agent is easy. Building a team that works across platforms is brutal. AgenTopology lets you define it once — then scaffold to Claude Code, OpenClaw, Codex, Cursor, and more.

View on GitHub
dev-pipeline : pipeline, fan-out, human-gatedesign-review[revise, max 2]intakeactionresearchersonnet · phase 1writersonnet · phase 2revieweropus · phase 3quality-checkgatearchitectsonnetsecurity-leadhaikumoderatorhumanAgentActionGateGroupHumanFlow
Write once. Scaffold anywhere.
Claude CodeOpenClawCodexCursorGemini CLICopilotKiro
The problem

The multi-agent configuration
nightmare

Platform lock-in

You build an agent team in Claude Code — 15 files, custom prompts, hooks, MCP servers. It works. Then you need it in OpenClaw. Or Cursor. Completely different config formats. You rewrite everything from scratch.

Invisible architecture

Your topology is scattered across AGENT.md files, SKILL.md files, settings.json, and scripts in nested directories. There's no single picture. A new team member has to reverse-engineer the entire system from config files.

No quality guarantees

There's no standard way to enforce gates between agents. No validation before deploy. No way to know if your agent wiring is broken until it fails at runtime. You test by running and hoping.

How it works

Three steps to a running
multi-agent system

01

Define

Write your agent team in a single .at file — agents, tools, flows, gates, groups. Or start from a template.

02

Validate & Visualize

29 built-in rules catch broken wiring before you deploy. See your topology as an interactive graph.

03

Scaffold

One command generates platform-native configs. Claude Code, OpenClaw, Codex, Cursor — same topology, any target.

The .at format

Human-readable agent
architecture

A declarative topology file that describes agents, their capabilities, and how they connect. No boilerplate.

dev-team.at
topology dev-team : [pipeline, fan-out, debate] {
  orchestrator { model: opus, handles: [intake] }

  agent researcher { model: sonnet, tools: [Read, WebSearch] }
  agent writer     { model: sonnet, tools: [Read, Write] }
  agent reviewer   {
    model: opus
    outputs: { verdict: approve | revise }
  }

  group design-review {
    members: [architect, security-lead]
    speaker-selection: "round-robin"
    max-rounds: 3
  }

  gates {
    gate quality-check {
      after: reviewer
      run: "scripts/lint.sh"
      on-fail: halt
    }
  }

  flow {
    intake -> researcher -> writer -> reviewer
    reviewer -> writer [when verdict == revise, max 2]
    reviewer -> design-review [when verdict == approve]
  }
}
design-review[revise, max 2]intakeresearchersonnet · tools: Read, WebSearchwritersonnet · Read, Writerevieweropus · verdictarchitectsonnetsecurity-leadhaikuquality-checkgate · on-fail: halt
Examples

See it in action

Real-world topologies you can explore, download, and scaffold in seconds.

Open source

An open standard for
agent architecture

AgenTopology is Apache 2.0. Your .at files work everywhere — no vendor lock-in, no proprietary formats.

0
Stars
0
Forks
1
Contributors
Pricing

Start building for free

Everything you need to design and scaffold agent systems.

Free

Open Source

Free forever
  • .at parser and validator
  • CLI scaffolding to all frameworks
  • All templates included
  • Community support
  • Apache 2.0 license
Get Started
Coming Soon

Pro

For teams shipping agents at scale
  • Visual topology designer
  • Team collaboration
  • Private template library
  • CI/CD integrations
  • Priority support