Semcom

The Semantic Compiler — from human intent to machine execution, no code required.

$ semcom compile "create a web server on port 8080"
Step 1: Refining program intent...
Step 2: Generating LLVM IR program... validation PASSED
Step 3: Building Docker image (AOT compile)...
Step 4: Running integration tests... 3/3 passed

$ curl localhost:8080
<!DOCTYPE html>...

This page is served by a static Linux binary compiled from LLVM IR that was generated by AI from a natural language prompt. The binary has zero dependencies — no libc, no runtime, no dynamic linker. It talks directly to the Linux kernel via raw syscalls from a FROM scratch Docker container.

24pipeline runs

373test cases

95.7%pass rate

0dependencies

The Whitepaper — The Semantic Compiler: Eliminating Code as the Intermediary Build Log — Generation Zero proof of concept results

How It Works

Intent Layer — Natural language is refined into a structured specification via Claude API
Meaning Compiler — Claude generates LLVM IR from the specification with a validate-repair loop
Executor — llvmlite JIT-compiles the IR and runs it against test cases (functions) or Docker AOT-compiles it (programs)
Semantic Bridge — Claude reads the IR and test results, produces a behavioral model of what the system actually does
Alignment Engine — Compares intent against behavior using deterministic and semantic tracks

What Generation Zero Proves

AI can translate natural language intent directly into correct LLVM IR
The closed semantic loop (intent → compile → execute → analyze → align) works in practice
Behavioral repair is possible: when tests fail, the bridge diagnoses the issue and the compiler fixes the IR
Pattern distillation captures what works and improves future generation quality
LLVM IR is a viable deployment target, not just a JIT intermediary — this server proves it

About

Lane Thompson is a software engineer based in Ocala, Florida. He wrote his first program in QBasic in middle school and has been building things with computers ever since. Semcom is his exploration of what happens when you remove code as the intermediary between human intent and machine execution — and let AI talk directly to the hardware.