CircuitForge AI
ActiveAn AI-assisted electronics design workspace: describe a project in plain language, get a real circuit — then simulate it, validate it, fix it, and export the build docs.
Last updated Jul 16, 2026
The problem
Getting from a hardware idea to a circuit you trust means bouncing between schematic tools, datasheets, and forum threads. Beginners in particular can't tell whether a design will brown out, backfeed through USB, or exceed a part's absolute maximum ratings until they've already built it and let the smoke out.
What’s being built
A design workspace backed by a real netlist, not a drawing. Projects, parts, and nets persist in Postgres; a DC operating-point solver propagates voltages from sources through chargers, switches, diodes, and regulators with load-dependent dropout, then rolls up currents and integrates battery discharge to drive live telemetry during simulation. A rules engine computes findings — overvoltage against absolute maximums, supply undervoltage, USB backfeed topology, I2C address conflicts and missing pull-ups, regulator headroom, switch contact ratings against startup surge, missing bulk capacitance — and auto-fixes mutate the netlist for real before the rules re-run. An AI engineering assistant works alongside the whole thing.
Latest update
The parts library carries behavioral models — supply ranges, absolute maximums, per-mode current draw, regulator dropout curves, battery capacity — so validation findings come from the real design rather than a checklist.
Current milestone
Simulation, validation, and auto-fix loop running against a live parts library with behavioral models
Tech stack
Follow this project
Get notified when there's a meaningful update — no noise, no marketing.
Confirm via email. Unsubscribe any time. No spam — ever.