Bridge between LLM-Agent and Cadence Virtuoso. A new infrastructure for Agentic Analog and Mixed-Signal Design.
Overview
A new infrastructure for . LLM Agents drive Cadence Virtuoso instances — locally or remotely — turning tedious handcrafting into automated design flows. — Control across Schematic, Layout, Maestro, and Spectre. - : execute inline SKILL, load .il files, or use Python APIs - : schematic editing, layout generation, simulation setup (Maestro), and standalone Spectre with PSF parsing — Multi-server, multi-session, built for distributed design clusters. - Multi-profile SSH: connect to N design servers, each with independent tunnel - Run parallel simulations across servers and accounts - Verified across macOS, Windows, and Linux — Built for coding agents (Claude Code, Cursor, etc.) to drive Virtuoso. - CLI-first: virtuoso-bridge start/status/restart, no GUI needed - Ships with pre-defined agent skill files (skills/) — the agent knows how to use the bridge immediately - Optimized for high-frequency agent interactions with persistent SSH tunnels , read AGENTS.
README
virtuoso-bridge.tokenzhang.com · Guide · 中文文档
Let AI agents drive Cadence Virtuoso and Spectre — locally or on a remote server — from Python, SKILL or the command line.
git clone https://github.com/Arcadia-1/virtuoso-bridge-lite.git
cd virtuoso-bridge-lite
uv venv .venv && source .venv/bin/activate
uv pip install -e .
virtuoso-bridge init user@host # write ~/.virtuoso-bridge/.env
virtuoso-bridge start # prepare fixed-port setup and the SSH tunnel
# Paste the printed load("...") into the intended idle CIW.
virtuoso-bridge status
Setup, configuration, CLI reference and architecture are in the
guide. Coding agents start with AGENTS.md.
Automatic sessions are opt-in: set VB_AUTO_SESSIONS=1 and optionally
VB_REMOTE_WORK_DIR=/projects/example to discover a registered daemon using an
OS-assigned port. A working directory alone keeps the legacy fixed-port mode.
Star History
Citation
If you use virtuoso-bridge in academic work, please cite:
@article{zhang2025virtuosobridge,
title = {Virtuoso-Bridge: An Agent-Native Bridge for Remote Analog and Mixed-Signal Design Automation},
author = {Zhang, Zhishuai and Li, Xintian and Sun, Nan and Jie, Lu},
year = {2025}
}
Author affiliation: Tsinghua University (Zhishuai Zhang is now at ETH Zurich).
Contact: [email protected]
Personal homepage: tokenzhang.com
Recommended Tools
Try a different keyword or remove a filter.
Install
npx skillfish add arcadia-1/virtuoso-bridge-lite