🤖 ⚡ ⚔️

Bring Your Own Agent

JEBAT's adapter system accepts any AI agent. Connect your custom agent to our memory, skills, and orchestration system.

Integration Steps

1

Create Your Agent

Any AI agent that can communicate via REST API or WebSocket works. Python, Node.js, Go — doesn't matter.

from fastapi import FastAPI
app = FastAPI()

@app.post('/chat')
async def chat(message: str):
    return {'response': process(message)}
2

Register with JEBAT Gateway

Point your agent to the jebat-gateway on port 18789. Register as a worker with a role and skill set.

POST /gateway/register
{
  "name": "my-agent",
  "role": "tukang",
  "skills": ["fullstack", "database"],
  "endpoint": "http://localhost:3001/chat"
}
3

JEBAT Routes Work to You

When a task matches your skills, Panglima (the orchestrator) sends it your way. You execute and stream results back.

POST /gateway/task
{
  "task": "Build a user CRUD API",
  "context": {...memory...},
  "thinking_mode": "deliberate"
}
4

Results Flow Back

Your results are streamed back through the gateway to wherever the user is — web UI, CLI, IDE, or Agent Town.

// Stream tool calls and results
ws.send(JSON.stringify({
  type: 'tool_use',
  tool: 'run_code',
  status: 'completed',
  result: '200 OK'
}))

Supported Protocols

Pick whichever your agent already supports.

REST API

Easy

Any HTTP endpoint

WebSocket

Medium

Real-time streaming

MCP Protocol

Medium

Model Context Protocol

OpenAI Compatible

Easy

Chat completions format

Ready to Connect?

Start Onboarding →View Source