Installing Waveflow

Waveflow requires Python 3.10 or newer. Pick the path that fits you:

  • User Setup — use Waveflow as a library in your own project. pip install from GitHub into a virtual environment; no clone needed. Start here if you just want to build designs with Waveflow.
  • Developer Setup — modify Waveflow, contribute changes, or run the bundled examples and tests. Clone the repository and install in editable mode with the development tools.

Either path gives you the full Python simulation flow with no other tools installed. To synthesize designs or run RTL-level simulations you additionally need AMD Vitis and Vivado 2025.1 or newer:

  • Connecting Vitis and Vivado — point Waveflow at your AMD installation with WAVEFLOW_VITIS_PATH, and confirm it with the bundled test_amd_tools command.

The optional AI-assistant integrations — MCP server, OpenAI-backed semantic example search, and the VS Code extension — are covered separately under AI Tooling. None of them are required to use or develop Waveflow.

Table of contents

  • User Setup
  • Developer Setup
  • Connecting Vitis and Vivado - Waveflow runs Python simulations with no AMD tools installed; synthesis and RTL-level simulation need Vitis and Vivado 2025.1 or newer. Point Waveflow at them by setting WAVEFLOW_VITIS_PATH to the directory holding your AMD installs — Vivado is found automatically beside Vitis — then confirm the setup with the bundled test_amd_tools command.