Overview

Heroku pioneered the PaaS model. RaidFrame builds on that foundation with modern infrastructure: Docker containers, auto-scaling, managed databases with branching, and AI-native tooling. Migration is straightforward if you have a Dockerfile or Buildpack-compatible app.

Step 1: Install the CLI

curl -fsSL https://get.raidframe.com | sh
rf auth login

Step 2: Export from Heroku

Environment Variables

heroku config --json --app my-heroku-app > heroku-env.json

Database

heroku pg:backups:capture --app my-heroku-app
heroku pg:backups:download --app my-heroku-app
# Downloads latest.dump

Procfile

Your existing Procfile maps directly to RaidFrame services:

web: node server.js
worker: node worker.js
clock: node scheduler.js

Step 3: Initialize on RaidFrame

rf init

Create raidframe.yaml from Procfile

services:
  web:
    type: web
    command: node server.js
    port: 3000
    scaling:
      min: 2
      max: 10

  worker:
    type: worker
    command: node worker.js

  clock:
    type: cron
    command: node scheduler.js
    schedule: "* * * * *"

Import Environment Variables

rf env push --file heroku-env.json

Step 4: Add Databases

# Provision
rf add postgres
rf add redis

# Import Heroku database
rf db import main latest.dump --format custom

Step 5: Deploy

rf deploy

Heroku → RaidFrame Mapping

Heroku	RaidFrame
`git push heroku main`	`rf deploy`
`heroku logs --tail`	`rf logs`
`heroku config:set`	`rf env set`
`heroku run bash`	`rf ssh web`
`heroku run rails console`	`rf exec web "rails console"`
`heroku pg:psql`	`rf db connect main`
`heroku addons:create`	`rf add postgres/redis`
`heroku ps:scale web=3`	`rf services scale web --min 3`
`heroku releases:rollback`	`rf deployments rollback`
`heroku pipelines:promote`	`rf deployments promote staging --to production`
Procfile	raidframe.yaml
Buildpacks	Dockerfile or auto-detect

What You Gain

Auto-scaling — Heroku requires manual ps:scale. RaidFrame scales automatically.
Faster deploys — Docker layer caching means ~10s deploys vs Heroku's 60-120s.
No slug size limit — Heroku's 500MB slug limit doesn't exist.
No dyno sleeping — Free tier idles after 15 minutes but wakes in under 3s (vs Heroku's 30s+ cold start).
Built-in cron — no clock process hack needed.
Database branching — fork your database for testing.
Preview environments — auto-deploy per pull request.
50-70% lower cost — comparable Heroku setup costs 2-3x more.

Heroku Add-ons → RaidFrame Built-ins

Heroku Add-on	RaidFrame
Heroku Postgres	`rf add postgres`
Heroku Redis	`rf add redis`
Heroku Scheduler	`rf cron add`
Papertrail	Built-in (rf logs)
New Relic	Built-in (rf metrics)
SendGrid	Built-in (rf add email)
Cloudinary	Built-in (rf add storage)
Bonsai (Elasticsearch)	`rf add search`
CloudAMQP	`rf add queue`

Most Heroku add-ons have a built-in equivalent on RaidFrame. Fewer providers, fewer bills.

Common Issues

PORT Environment Variable

Heroku sets PORT dynamically. On RaidFrame, set it in your config:

services:
  web:
    port: 3000

Your app should still read process.env.PORT — RaidFrame injects it automatically.

Buildpacks

RaidFrame supports Docker and auto-detection. If you relied on custom Heroku buildpacks, create a Dockerfile instead. For standard stacks (Node, Python, Ruby, Go), auto-detection works out of the box.

Migrate from Heroku