Kubernetes-native functions platform with sync and async execution, Git-driven updates, and first-class multi-language runtimes.
Bottom line:
- No vendor lock-in.
- No language lock-in.
LiteFunctions is built around first principles:
- Keep the control plane in your cluster.
- Keep source-of-truth in your repo.
- Keep runtime choices open.
2026-02-13.00-37-42_2x_muted.mp4
- User calls an endpoint from the Portal or an external client.
- Ingestor receives the request and asks Operator to ensure the function/runtime is active.
- Operator reconciles Function CRDs and ensures deployment/service readiness.
- For sync execution, Ingestor proxies HTTP to runtime services.
- For async execution, Ingestor publishes execution events over NATS subjects.
- Runtime workers consume events, execute function code, and publish responses/results.
- Dynamic language runtimes (Python/TS/Lua) refresh function code from VCS using hook events instead of per-function image builds.
- Portal tracks runs, status, and endpoint behavior via APIs and live updates.
helm install litefunctions oci://registry-1.docker.io/ashupednekar535/litefunctions- Teams running Kubernetes who want function-style deployment without moving to a managed FaaS vendor.
- Platform engineers who need predictable control over network, storage, auth, and CI/CD.
- Product teams that want one platform for multiple languages and execution styles.
- Self-hosters and enterprises that care about portability, sovereignty, and auditability.
Current platform features:
- Kubernetes-native function lifecycle via CRDs and operator reconciliation.
- Multi-language runtimes: Go, Rust, Python, TypeScript (Bun), Lua.
- Sync HTTP execution and async pub/sub execution.
- Dynamic runtime refresh for Python/TS/Lua via VCS hook events.
- Git-integrated workflow support (Gitea/GitHub workflow templates).
- Auto endpoint provisioning and endpoint management from Portal.
- Build/action visibility in Portal with live status updates.
- Runtime activation + keep-warm/deprovision lifecycle management.
Planned next:
- Authorization policy engine for endpoints (authz policies).
- Richer authn/authz integration across HTTP/WS/SSE surfaces.
- More granular routing and traffic controls.
- Enhanced observability (structured traces/metrics per function).
- Runtime hardening and policy guardrails for multi-tenant environments.
