get started
Install the EML package. Track Forge honestly.
The public install today is monogate. It is the EML math, routing, optimization, search, and ML research package. The Forge compiler CLI and multi-target backend list are not public installables yet.
step 1 — install today
Install the public EML research package
Use this for the capabilities users can verify now: the EML operator, elementary functions, BEST routing, symbolic optimization, search experiments, and PyTorch layers.
pip install monogate
step 2 — roadmap
Join the Forge compiler access list
Public today
The EML research package. Good for learning, experiments, optimizer work, and ML activation research.
- EML operator arithmetic
- Elementary functions through EML/BEST routing
- Symbolic optimizer with per-operation savings
- MCTS search for exploratory expression fitting
- PyTorch layers for activation experiments
Forge roadmap
Compiler targets, hardware-oriented codegen, and proof exports are being developed behind private evidence gates.
- No public monogate-forge package yet
- No shipped public claim for C/Rust/Verilog/MATLAB/zkproof targets yet
- Private evidence packets exist for parts of the Forge/eFrog path
- Public access should open only when reproducible examples ship