Deployment

Expose your agents and workflows over HTTP with the deployer. It builds a Hono app that runs on Node, Bun, and the edge.

The HTTP server

import { serve } from '@hono/node-server'
import { createHonoServer } from '@redtuma/deployer'

serve({ fetch: createHonoServer(redtuma).fetch, port: 3000 })

Registered components are exposed as JSON routes:

• GET /api/agents — list agents
• POST /api/agents/:id/generate — run an agent
• POST /api/agents/:id/stream — stream a response
• POST /api/workflows/:id/run and /resume — run and resume workflows

Edge runtimes

On Workers or other edge runtimes, export the fetch handler directly — Hono runs natively.

import { toFetchHandler } from '@redtuma/deployer'

export default { fetch: toFetchHandler(redtuma) }

Durable workflow state

When the Redtuma instance has a Store configured, suspended workflow runs are persisted and can be resumed on any instance — so human-in-the-loop flows survive restarts and work across stateless isolates.