Skip to content
DUUMBI Docs

Introduction

DUUMBI is an AI-first semantic graph compiler. Instead of storing programs as text files, DUUMBI represents program logic as typed JSON-LD semantic graphs and compiles them directly to native binaries via the Cranelift backend.

Core thesis

Section titled “Core thesis”

Programs should be structured, verifiable data — not strings of characters that require parsing.

In DUUMBI, a function isn’t a sequence of characters. It’s a graph node with typed edges to its parameters, body blocks, and operations. This means:

  • No syntax errors — the graph structure is always valid by construction
  • AI generates programs directly — LLMs produce structured graph patches, not text that needs parsing
  • Every mutation is validated — schema validation and type checking happen at the graph level

How it works

Section titled “How it works”
.jsonld files → parse → StableGraph → validate → Cranelift IR → native binary
  1. JSON-LD source — Programs are stored as .jsonld files using the duumbi: namespace
  2. Semantic graph — Parsed into a petgraph::StableGraph — the single source of truth
  3. Validation — Schema validation, type checking, ownership verification
  4. Cranelift compilation — Graph nodes lowered to Cranelift IR, one function per subgraph
  5. Native binary — Object file linked with C runtime via cc

Key features

Section titled “Key features”
  • Semantic Graph IR — petgraph-backed StableGraph as the program representation
  • AI-native mutations — Multi-provider support (Anthropic, OpenAI, Grok, OpenRouter)
  • Cranelift backend — Fast compilation to x86-64 / AArch64, no LLVM dependency
  • Intent-driven development — Describe what you want, DUUMBI builds it
  • Ownership model — Rust-inspired borrow checking at the graph level
  • Module registry — Publish and install modules with SemVer resolution

Next steps

Section titled “Next steps”