Skip to main content
You write the agent logic. Reminix handles everything else — deploy your handlers in TypeScript or Python and get production REST APIs for agents and MCP servers for tools, with streaming, auth, connections, and monitoring built in.

​
Define an agent

import { agent, serve } from "@reminix/runtime"

const bot = agent("support-bot", {
  type: "chat",
  description: "Customer support agent",
  handler: async (input) => {
    return "How can I help you today?"
  },
})

serve({ agents: [bot] })

​
Call it with the SDK

import Reminix from "@reminix/sdk"

const client = new Reminix({ apiKey: "your-api-key" })

const response = await client.agents.chat("support-bot", {
  messages: [{ role: "user", content: "Help me reset my password" }],
})

​
Or use the CLI

Invoke agents and call tools directly from the terminal — no client code needed. 
# Invoke an agent
reminix agent invoke support-bot --prompt "How do I reset my password?"

# Stream the response
reminix agent invoke support-bot --prompt "Summarize this report" --stream

# Call a tool
reminix tool call lookup_customer -i '{"email": "user@example.com"}'
The CLI also works inside AI agents (Claude Code, Cursor), CI/CD pipelines, and shell scripts. See Scripting and CI/CD for more.

​
Or use your existing framework

Already using Vercel AI SDK, LangChain, OpenAI, or Anthropic? Wrap your existing code and deploy it as a Reminix agent with a few lines. 
import { serve } from "@reminix/runtime"
import { VercelAIChatAgent } from "@reminix/vercel-ai"
import { openai } from "@ai-sdk/openai"

const bot = new VercelAIChatAgent(openai("gpt-4o"), {
  name: "support-bot",
  instructions: "You are a helpful support agent.",
})

serve({ agents: [bot] })

​
Key capabilities

  • Agents → REST APIs — Tasks (one-shot), Conversations (multi-turn), Workflows (pause/resume)
  • Tools → MCP servers — Define tools served via the Model Context Protocol, discoverable by Claude, Cursor, Windsurf, and any MCP client
  • Built-in LLM access — Every project includes LLM access out of the box. Bring your own keys when you’re ready for production
  • OAuth & Connections — Managed OAuth for 20+ services. Your code gets a valid token — Reminix handles the flow, storage, and refresh
  • Streaming — Real-time Server-Sent Events for progressive responses
  • Framework integrations — Drop-in support for OpenAI, Anthropic, LangChain, Vercel AI SDK, and Google

​
Next steps

Quickstart

Get up and running in under 5 minutes.

Agents

Learn about agent types and how to define them.

Tools & MCP

Define tools and serve them via MCP.