Find an agent. Copy the source. Own the code.

One command. The code is yours.

Three steps from registry to running in your own stack — no lock-in, no hidden runtime.

1. 1

   Add it

   Run one command. The CLI copies the full agent source into ./agents/<id>/ in your repo. From here, the code is yours — read it, edit it, commit it.

   npx agentskit add research
2. 2

   Wire it

   It is plain TypeScript in your project. Import the factory and pass any adapter — OpenAI, Anthropic, or any of 140+ catalog providers. Swap the model whenever you want.

   import { openai } from '@agentskit/adapters'
   import { createResearchAgent } from './agents/research/agent'
   
   const agent = createResearchAgent({
     adapter: openai({ apiKey: process.env.OPENAI_API_KEY! }),
   })
3. 3

   Run it anywhere

   Call agent.run() from Node, a serverless function, an edge runtime, or your terminal. It is your code on your infrastructure — nothing phones home to this registry.

   const { content } = await agent.run(
     'What changed in the EU AI Act in 2025?',
   )
   console.log(content)

   Your code · your stack

   $ npx tsx research.ts

   running research agent…