defineRoute({ method, path, handler })
Typed edge routes
Write a function, export a route. Forgepoint infers request and response types end to end, so a schema change in your handler surfaces as a compile error, not a 2 a.m. page.
Edge API runtime · v4 GA
Ship the endpoint. We handle the edge.
Forgepoint compiles your TypeScript functions to a globally distributed edge runtime — typed routes, sub-30ms cold starts, and zero infrastructure to babysit. Deploy with one command and read the logs before the request finishes.
$
npm i -g @forgepoint/cli && forge deploy- ✓No credit card required
- ✓310 edge regions
- ✓99.99% uptime SLA
Powering teams that ship to the edge
HeliographNimbus LabsQuantaryBeaconDriftwoodNorthgate
Capabilities
The runtime gets out of your way.
Typed routes, scale-to-zero billing, and a global edge — so the only thing you maintain is application logic.
cold start p50 · 27ms
Scale to zero
Idle functions cost nothing. The runtime spins a fresh isolate in under 30ms on the first request and tears it down when traffic stops — no warm-pool tax, no minimum spend.
region: "auto" · 310 PoPs
Global by default
Every deploy lands in 310 regions at once. Requests resolve from the nearest point of presence, and stateful data pins to the regions you choose with a single config flag.
forge logs --tail --trace
Streaming logs
Tail structured logs and traces in real time from the CLI or dashboard. Every request carries a trace ID you can grep, replay, and pipe straight into your existing observability stack.
ts · py · go · rust
SDKs for every runtime
First-class clients for TypeScript, Python, Go, and Rust — generated from your route types, versioned with your deploys, and published to your registry on every release.
forge rollback <build-sha>
Instant rollback
Each deploy is an immutable, content-addressed snapshot. Promote, alias, or roll back to any previous version in one command — traffic shifts atomically with zero dropped requests.
27ms
Median cold start, p50
310
Edge regions, every deploy
8.4B
Requests served each day
99.99%
Uptime over the last 12 months
Pricing
Pay for requests, not for seats.
A generous free tier, transparent usage pricing, and no per-developer license to renegotiate every quarter.
Hobby
$0forever
For prototypes, side projects, and your first edge function.
- ✓1M requests / month
- ✓100 GB-s compute
- ✓3 projects
- ✓Community Discord support
- ✓24-hour log retention
Most popular
Pro
$20per month + usage
For teams shipping production traffic on the edge.
- ✓25M requests included, then $0.30/M
- ✓Unlimited projects & environments
- ✓Custom domains & wildcard routes
- ✓14-day log & trace retention
- ✓Priority email support
Scale
Customannual agreement
For platforms with compliance, volume, and SLA requirements.
- ✓Volume request pricing
- ✓Region pinning & private networking
- ✓SSO, SCIM & audit log streaming
- ✓99.99% uptime SLA
- ✓Dedicated solutions engineer
FAQ
Questions developers actually ask.
Runtimes, billing, data residency, rollbacks — answered without the marketing fog.
What runtimes can I deploy?+
Forgepoint runs a V8 isolate runtime with full Web Standard APIs — fetch, streams, crypto, and Web Sockets. Author handlers in TypeScript or JavaScript today; WASM modules compiled from Rust and Go are supported on Pro and Scale.
How does scale-to-zero billing work?+
You are billed per request plus GB-seconds of compute while a handler is actually executing. Idle functions accrue zero charges, and the dashboard shows live spend with a soft cap you can set per project.
Can I pin data to a specific region?+
Yes. Set region: "fra1" (or any region code) on a route, or define a residency policy at the project level. Stateful primitives — KV, queues, and the edge cache — honor the same pin so your data never leaves the boundary you choose.
How do rollbacks work?+
Every deploy is an immutable, content-addressed build. forge rollback <build-sha> shifts 100% of traffic to that build atomically. You can also alias builds to environments and run weighted canaries between two versions.
Do you offer an on-prem or self-hosted option?+
The managed edge is our primary product, but Scale customers can run the Forgepoint runtime inside their own VPC or a private region cluster. Reach out to sales for a deployment architecture review.
$forge deploy
From localhost to 310 regions in one command.
Install the CLI, point it at a route, and watch it go live worldwide before your coffee cools. Your first million requests every month are on us.