Kedge: Globally distributed serverless platform with forkable VM snapshots

What it does

Kedge is a cloud platform for deploying websites, apps, and databases. Users push code or content via SSH, Git, or container image and get a live HTTPS URL. The platform handles database provisioning, CDN distribution, and global routing across ten regions. Apps scale to zero when idle. A persistent "dev machine" feature lets developers maintain checkouts, tools, and build caches between sessions, optionally paired with a coding agent.

Who it is for

Kedge targets developers and small teams building static sites, web applications, functions, and containerized services. The dev machine feature appeals to those who want persistent development environments. The emphasis on SSH-based deployment and minimal configuration suggests experienced developers. The coding agent support indicates interest from users exploring AI-assisted development.

Pricing

Kedge bills on actual resource consumption: CPU at $15 per vCPU-month, memory at $5 per GB-month, and storage at $0.05 per GB-month. Egress costs $0.01 per GB. A free tier provides $5 monthly in credits, sufficient for small websites or dev boxes. The platform uses credits-based billing and measures usage per second on consumed resources, not allocated capacity. Idle apps incur only storage costs.

How it stands out

Kedge emphasizes hardware isolation—every instance, handler, sandbox, and CI job gets its own Linux kernel rather than sharing. The dev machine concept blends persistent development environments with serverless deployment. Per-second metering of actual consumption differs from capacity-based pricing common in other serverless platforms. The SQLite replication model and 1.2 ms resume from idle are specific technical advantages claimed. The global distribution across ten regions is included by default rather than as a premium feature.

What a founder should check

A competitor should investigate how quickly Fly.io, AWS Lambda, Vercel, and Railway have evolved their own cost structures and feature sets. The dev machine model is relatively novel—verify whether existing platforms can replicate this workflow and how sticky it becomes for users. Examine whether the per-second metering and zero-scale-down costs materially beat incumbents on typical workloads; published benchmarks comparing total monthly bills across use cases would clarify the real advantage. Finally, understand what prevents lock-in: whether users can easily export their databases, cached build artifacts, and application state if they switch platforms.

Thinking of building something like this?

Every launch here is a competitor to somebody's idea. If yours is close, check it against the market before you build: the Full Check names the rivals, the prices and the gaps.