build once.
reuse everywhere.
bldr is a distributed build system where every artifact is a content-addressed block. identical inputs always resolve to the same output — so nothing is ever built, or transferred, twice. on top of that core sits a corporate-grade CI/CD platform: triggered builds, deployment pipelines, secrets and audit.
a pipeline that is a pure function
a workspace tracks upstream sources, functions turn them into content-addressed outputs, and deployments push those outputs to the world. the build controller sits in the middle, and every intermediate result is cached by its hash.
- › git · branch → sha
- › oci · tag → digest
- › s3 / http / path
- › fs.merge · fs.diff
- › pod.dag · oci.rootfs
- › blob.untar · http.get
- › k8s.apply
- › ssh.scp · ssh.exec
- › aws.s3 · oci.push
content-addressed, end to end
content-addressed everything
sources, layers and outputs are addressed by a blake3 hash. identical inputs share one object — the cache is exact, not heuristic.
reproducible sandboxes
every step runs in an isolated, rootless pod with a content-addressed overlay rootfs. no host leakage, deterministic results.
distributed cache
objects are pulled from and pushed to peers directly. a cold machine warms itself from the fleet instead of rebuilding.
typed build definitions
build logic is ordinary code with a small sdk — types, imports and composition, not a bespoke dsl.
pure builds
a build is a pure function from input cids to output cids. same inputs, same output — from cache, a peer, or execution, in that order.
one daemon, many surfaces
drive it from the cli, the grpc api, or the console. a single daemon owns the object store and schedules the work.
open-source engine, managed orchestration
the build system is open source and runs anywhere. bldr cloud adds the orchestration a corporate ci/cd system needs — without ever locking away the core.
- ✓the build system & dag runner
- ✓content-addressed object store
- ✓the function library (fs, blob, pod, oci…)
- ✓distributed peer cache
- ✓the bldr cli & daemon
- ✦automatically triggered builds on upstream change
- ✦multi-stage deployment pipelines & promotions
- ✦fleet-scale secrets, rbac & audit logs
- ✦observability: task trees, timings, cache stats
- ✦hosted build fleet & shared team cache
rebuild the moment a git sha or image digest moves.
chain builds and deployments with gates and promotions.
aes-256-gcm secrets, scoped per tenant and integration.
every action attributed, every deploy accountable.
stop rebuilding what you already built
spin up the open-source daemon in minutes, or start on bldr cloud and let the platform trigger, cache and ship for you.