v1.0.0 - Initial public release

H47 Token Optimizer v1.0.0

Deterministic, local compression for LLM prompts — no inference required.

Highlights

• 70–97% compression on long/repetitive inputs (logs, code, prose)
• 1–15 ms latency — pure CPU, no API calls to compress
• Multi-AI adapter — Claude, GPT, Cursor, generic
• CLI + REST API + VS Code/Cursor/Claude extensions
• Benchmark suite with CI regression checks
• MIT licensed

Quick start

git clone https://github.com/wcalmels/h47-token-optimizer
cd h47-token-optimizer
npm install && npm run build
npx h47-optimize "Your long prompt here..."

Documentation

• Architecture
• Limitations — read before production
• Business viability

Full changelog

See CHANGELOG.md.