Cloud FPGA infrastructure for hardware-reasoning agents.
Manhattan Reasoning runs the open-source Yosys + nextpnr toolchain on managed cloud FPGAs — synthesis, place-and-route, and on-device verification in seconds, not hours. It's built for agents learning hardware design through reinforcement learning with verifiable rewards: every submission is checked against a formal spec and on-device behavior for correctness, timing closure (Fmax), and resource usage (LUTs, FFs, DSPs).
A graded benchmark of classic NP-complete problems, reframed as hardware-design tasks. Agents reason from a formal specification to a synthesizable Verilog accelerator, then earn verifiable rewards across three axes — correctness, speed, and area.
| Repo | What it is |
|---|---|
| manhattan-reasoning-gym | SDK + CLI for local FPGA synth/PnR, sandboxed agents, and cloud silicon |
| mrg-sandbox | Open FPGA build environment: pinned toolchain image (Yosys/nextpnr/LiteX) + ECP5 SoC firmware |
| manhattanreasoning-docs | User documentation for manhattanreasoning.com |
| manhattanreasoning-web | Marketing landing page |
- 🌐 Website: manhattanreasoning.com
- 📚 Docs: docs.manhattanreasoning.com
- ✉️ Contact: hello@manhattanreasoning.com