Opico

Interested in Opico?

Contact the developer on LinkedIn or by email

Under Active Development - See Roadmap

High-Level Features:

Netlist-driven workflow

A simple interface to build complex simulation pipelines for PIC design.

Draw a study with blocks and wires.
Hierarchical sheets: build big studies from small ones.
Reuse one pipeline across devices and runs instead of rewriting scripts.
Keep everything organized and clear with drawing elements.

State-of-the-art design exploration

Answer complex design questions with efficient, powerful design-space exploration algorithms.

Bayesian, population-based and multi-objective optimization with mature open-source libraries.
Monte Carlo runs show how process variation spreads your results.
Convergence checks using the Grid Convergence Index method (GCI) so you can quantify the numerical precision of simulations.
Sensitivity analysis shows which parameter matters.
Combine strategies: run a Monte Carlo inside an optimization, or a sweep around the best point. Your creativity and hardware is the limit.

Open-source solver stack

Built-in tools from popular free solvers that cover most of PIC design.

Optical: femwell modes, fdtdx and beamz FDTD on the GPU.
Electrical and RF: DEVSIM junctions, Palace microwave lines, jaxfem heat.
Circuits: SAX and circulax.

Analytical models library

Fast analytical closed-form models and material data, ready to use.

A toolbox of multi-physics analytical models for common PIC applications.
A simulation-ready material database with rich physical properties (thermal, mechanical, electrical and more) and anisotropic effects.
Built-in, ready-to-use open-source photonic PDK on different material platforms (SiEPIC, Luxtelligence, ...).

GDSFactory integration

Define simulation geometry and circuit netlists with GDSFactory cross-sections and components.

Any GDSFactory component or cross-section can be imported as a block.
Live layout preview on the block.
The layout you tape out is the geometry you simulate.
Connect the inputs of parameterized GDSFactory functions to Opico exploration strategies.
Gmsh meshing back end.

Data persistence and rich visualization

A custom rendering engine for a rich library of complex and custom plots.

Plots update live during a run.
Export publication figures (PDF, pgf, SVG, PNG) and data (CSV, Parquet).
Every run is saved in an SQLite database: browse and compare past runs.
Don't lose simulation results ever again.

Hackable and extensible

A Python block connects Opico to any Python function; the block library grows with your own functions.

Wrap any Python function as a block.
Strategies are netlists too: open one and see how it works.
Create your own reusable netlist blocks.
Share your block library with your team.

Scalable simulation for HPC

Parallel runs and smart job scheduling make full use of your computing hardware.

Parallel workers run exploration strategies (optimization, sweeps, and more).
A result cache never runs the same point twice.
The same study runs on a laptop, and later on an HPC server.
Powerful tools to visualize resource usage, and accurate ETA estimates for large simulations.

MCP server integration

A light and powerful MCP server gives any AI full control for AI-assisted design.

FULLY OPTIONAL feature.
The AI can see and edit your sheets to help you build workflows.
Fully autonomous AI-assisted design of complex studies with Ralph loops.
Works with any AI provider.
Context-efficient, so smaller local models with limited context window work well too.

Lightweight Rust core

The core is written in Rust: fast, and light on memory.

Native, GPU-drawn app.
Light on memory = more room for simulation.
A solver crash never takes down the app.

Roadmap

Proof of Concept

Q4 2025 - Q2 2026

Proof of concept with a custom Python library.
Use and viability proven on industry projects.
Exploration strategies and solvers tested on real designs.
Proof of concept still being used as of today.

Development Build

Q3-Q4 2026

New Rust back-end and front-end architecture.
Proof of concept migration from Python.
Rust core and custom GPU graphics pipeline.
Proven Python solvers and analytical models carried over.

Alpha Version

End of Q4 2026

Internal alpha for close collaborators.
Close collaborators can experiment with Opico on their own designs.
Their valuable feedback will help work towards the public beta.

Beta Version

Q2 2027

Public beta release on GitHub.
Open source under GPL-3.0: free, forever.
Community can start contributing blocks, libraries and new features.