Show HN: Foundation, a different approach to software and AI
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- Ovasabi Foundation is pitched as a "full-stack application substrate for high-performance, event-driven systems" — an integrated toolkit that ships platform modules, scaffolds, enforcement checks, and documentation for bootstrapping and maintaining production codebases.
- Foundation enforces a seven-tier performance ladder with measured cost per plane: direct dispatch at 10–30 ns/op, binary frames at 20–80 ns/op, generated protobuf ~370 ns/op, gRPC 20–30 µs/op, and JSON ~30 µs/op, plus native FFI/SHM and browser/WASM planes.
- Foundation's core rule: "the fastest lane must not pay the cost of the compatibility lane," with regressions caught automatically before they land rather than caught in production.
- Hermes hotplane provides node-local, memory-bounded, indexed read models that project database mutations in real time to enable sub-microsecond operational reads, sitting alongside PostgreSQL for durable truth, Redis for coordination, Protocol Buffers for contracts, and Cap'n Proto for zero-copy boundaries.
- Foundation explicitly rejects low-discipline alternatives — "Not a no-code platform. Not zero-DevOps. Not for teams that want to move fast by cutting corners" — and ships dedicated AI documentation including AGENTS.md, an agent operating contract, and an AI threat model.
- Generated projects consume Foundation through package boundaries rather than raw imports, bootstrapped via
ovasabi.js initor./scripts/init-project.sh, with Makefile targets covering lint, test, rust checks, CI verification, practice-control matrices, and link validation.
Why it matters: For backend teams willing to accept Foundation's discipline requirements, the framework packages what most teams assemble piecemeal — tenant isolation from authenticated context, correlation-ID-linked observability, bounded workers with exponential backoff, and seven enforced performance budgets with pre-merge regression blocking — into a scaffolded project. It targets the gap between nanosecond hardware capability and the millisecond latencies of typical software stacks.
