Cloudflare platform coverage.
Keep your Worker code, Wrangler configuration, framework adapters, and bindings. See the current product status before deploying.
Works with your ecosystem
#1 Requires an external LLM API.#2 Requires an external sidecar.#3 Partial or work in progress; see documentation.
Deploy, run, and operate Workers.
OCD + WRANGLER
Keep your project. Change the target.
- Same wrangler.jsonc
- Immutable deployments and rollback
- Local or remote infrastructure
$ ocd status
$ ocd wrangler deploy --env production
✓ uploaded 12 modules
✓ deployment v42 active
✓ /api/* → v42WORKER RUNTIME
Standard Worker runtime.
- Fetch, Streams, Crypto, and WebSockets
- Standard module Worker syntax
- Isolates instead of containers
type Env = {
GREETING: string;
};
export default {
fetch(request: Request, env: Env) {
return Response.json({
message: env.GREETING,
path: new URL(request.url).pathname,
});
},
} satisfies ExportedHandler<Env>;STATE + SERVICES
Bindings through env.
- Familiar Cloudflare binding APIs
- Capabilities declared in Wrangler
- Local or S3-backed authority
type Env = {
CACHE: KVNamespace;
DB: D1Database;
BUCKET: R2Bucket;
JOBS: Queue<Job>;
};
const profile = await env.CACHE.get("user:42", "json");
const row = await env.DB
.prepare("SELECT * FROM jobs WHERE id = ?")
.bind(id)
.first();
await env.BUCKET.put(key, request.body);
await env.JOBS.send({ id });OCD
One local control plane.
- One binary and one managed service
- Supervised workerd runtime
- One-time Dashboard login
$ ocd setup --yes
✓ service configured
✓ runtime ready
$ ocd status
READY workerd supervised
STORAGE local authority
$ ocd dashboard
→ opening one-time loginA compact stack for self-hosted Workers.

//001Isolates, not containers.
Native workerd isolation without a container per request.

//002One binary. No sidecars.
Runtime, control plane, storage integration, and Dashboard ship together.

//003Authority stays local.
SQLite is authoritative. Deployments are immutable. Artifacts are content-addressed.
Self-host for free.
Everything you need to self-host:
- Apache-2.0 open-source platform
- Compatible runtime and product bindings
- Community support on GitHub
A managed open-compute experience:
- Hosted open-compute infrastructure
- Managed upgrades and backups
- Cloud dashboard and team access
Everything in Self-host, plus:
- Deployment and migration guidance
- Security and production-readiness review
- Direct engineering support
SEP 10, 2026
0.1.4
SEP 10, 2026
0.1.4 adds a complete Cloudflare Artifacts implementation for self-hosted open-compute installations. Operators can create namespaces and repositories, use Wrangler and Worker bindings, push and clone through Git Smart HTTP, and include repository data in the existing snapshot and restore lifecycle.
SEP 09, 2026
0.1.3
SEP 09, 2026
0.1.3 adds a project-native Wrangler workflow for teams running Cloudflare-compatible Workers on a self-hosted open-compute instance. A repository can now keep an ordinary wrangler.jsonc, select a named open-compute target, and use pinned Wrangler commands without copying control-plane URLs or credentials into project files.
SEP 08, 2026
0.1.2
SEP 08, 2026
0.1.2 is the first release with a complete day-to-day operator workflow for a single-machine open-compute installation. You can now initialize an instance, run it as an OS service, inspect it, open its Dashboard, upgrade it, and uninstall it safely.