Deployment

Deploy your autonomous agents to production. We support Edge Runtimes (Vercel, Cloudflare) and containerized environments (Docker, K8s).

Deployment Strategies#

Edge Runtime

Best for low-latency, stateless agents. Runs on Vercel or Cloudflare Workers.

Docker Container

Best for long-running tasks, cron jobs, and agents requiring heavy dependencies.

Akios Cloud

Fully managed control plane. Zero devops. (Enterprise only)

Deploy to Vercel#

Akios agents are compatible with the Vercel AI SDK and Next.js App Router.

Edge Compatibility

Ensure you use the edge runtime in your route handler configuration for maximum performance.

1. Create Route Handler

app/api/agent/route.ts

import { Agent } from '@akios/core'
import { StreamingTextResponse } from 'ai'

export const runtime = 'edge'

export async function POST(req: Request) {
  const { messages } = await req.json()
  
  const agent = new Agent({
    name: 'edge-assistant',
    model: 'gpt-4-turbo',
    systemPrompt: 'You are a helpful assistant.'
  })

  // Stream the response back to the client
  const stream = await agent.stream(messages.at(-1).content)
  return new StreamingTextResponse(stream)
}

2. Deploy

bash

vercel deploy

Deploy with Docker#

For background workers or custom infrastructure, use our official base image.

Dockerfile

FROM node:20-alpine

WORKDIR /app

COPY package*.json ./
RUN npm ci --only=production

COPY . .
RUN npm run build

# Start the agent listener
CMD ["npm", "start"]

Environment Variables

Ensure the following variables are set in your container runtime:

OPENAI_API_KEY
AKIOS_LICENSE_KEY (if using Enterprise features)
REDIS_URL (for memory persistence)