Move from Fly.io to RaidFrame — simpler networking, flat pricing, more built-in services.

Overview

Fly.io runs Docker containers on bare metal globally. RaidFrame does the same with a simpler mental model — no Fly Machines API, no fly.toml quirks, no surprise bills from idle machines you forgot about.

What Replaces What

Fly.io	RaidFrame
Fly Machines	Container services
`fly.toml`	`raidframe.yaml`
`fly deploy`	`rf deploy`
`fly postgres`	`rf add postgres`
Upstash Redis (Fly addon)	`rf add redis` (managed)
Tigris (Fly addon)	`rf add storage`
`fly logs`	`rf logs`
`fly scale`	`rf services scale`
`fly ssh console`	`rf ssh`
`fly secrets`	`rf secrets set`
Fly Proxy / Anycast	Built-in load balancer
Fly Volumes	Object storage or persistent volumes
Fly Extensions	Built-in services

Step 1: Export from Fly.io

fly.toml → raidframe.yaml

# fly.toml
app = "my-app"
primary_region = "iad"

[build]
  dockerfile = "Dockerfile"

[http_service]
  internal_port = 8080
  force_https = true
  auto_stop_machines = true
  auto_start_machines = true
  min_machines_running = 1

[[services]]
  internal_port = 8080
  protocol = "tcp"

  [[services.ports]]
    port = 443
    handlers = ["tls", "http"]

[env]
  NODE_ENV = "production"

Becomes:

# raidframe.yaml
project: my-app
region: us-east-1

services:
  web:
    type: web
    build:
      dockerfile: Dockerfile
    port: 8080
    scaling:
      min: 1
      max: 10
      scale_to_zero:
        idle_timeout: 15m

env:
  shared:
    NODE_ENV: production

Environment Variables

fly secrets list --json | jq -r '.[] | "\(.Name)=\(.Digest)"'
# Fly doesn't expose secret values — re-set them manually

rf secrets set DATABASE_URL=postgresql://... REDIS_URL=redis://... API_KEY=sk_live_...

Database

# Connect to Fly Postgres
fly postgres connect -a my-app-db

# From inside psql:
# \copy (SELECT * FROM users) TO '/tmp/users.csv' CSV HEADER;

# Or use pg_dump from outside
fly proxy 15432:5432 -a my-app-db &
pg_dump -h localhost -p 15432 -U postgres mydb > backup.sql

rf add postgres
rf db import main backup.sql

Volumes → Storage

Fly Volumes are local disk attached to machines. Move to object storage:

fly ssh console -a my-app
tar czf /tmp/data.tar.gz /data
exit

fly sftp get /tmp/data.tar.gz ./data.tar.gz -a my-app
tar xzf data.tar.gz

rf add storage
rf storage sync ./data/ s3://my-app-data/

Step 2: Deploy

rf deploy
rf domains add myapp.com

What You Gain

Feature	Fly.io	RaidFrame
Pricing model	Usage-based (CPU/RAM per second)	Flat per-service
Idle machine costs	You pay for forgotten machines	Scale-to-zero or predictable flat
Managed Postgres	Fly Postgres (self-managed on Fly)	Fully managed (backups, replicas, branching)
Managed Redis	Via Upstash addon ($)	Built-in (`rf add redis`)
Object storage	Via Tigris addon ($)	Built-in (`rf add storage`)
Message queues	Not built-in	Built-in
Search	Not built-in	Built-in
Email	Not built-in	Built-in
Database branching	Not available	Fork, test, merge
CI/CD	Not built-in (use GitHub Actions)	Built-in
Preview environments	Manual	Automatic per PR
Traffic mirroring	Not available	Mirror prod → staging
MCP integration	Not available	Full AI assistant support
Cost forecasting	Not available	Predict and optimize

Fly.io-Specific Issues

Fly Machines vs Containers

Fly Machines are Firecracker microVMs. RaidFrame runs Docker containers. Your Dockerfile works identically — the runtime difference is transparent to your app.

Fly Multi-Region

On Fly, you fly scale count 2 --region iad,cdg. On RaidFrame:

rf regions add eu-west-1

Database read replicas follow automatically if configured.

Fly Wireguard

If you use Fly's private WireGuard network for development, RaidFrame has the same:

rf vpn connect

Fly Proxy / Replay Header

Fly uses a fly-replay header for multi-region request routing. On RaidFrame, latency-based routing is automatic. Remove any fly-replay logic from your app.

Migrate from Fly.io