This product was not featured by Product Hunt yet. It will not be visible on their landing page and won't be ranked (cannot win product of the day regardless of upvotes).
Cascodex
An AI agent that verifies analog chips at every PVT corner
Analog circuits rarely fail at typical. They fail at the edges: fast silicon, a hot die, a low supply. Cascodex simulates all 45 PVT corners of your design in ngspice. When a spec fails, an AI agent traces it to the device causing it, edits the netlist and re-runs every corner before calling it fixed. On our benchmark it found verified fixes in 32 of 33 runs, at about $0.08 each. Corner maps, waveforms, shareable reports and CI included. Free on the open SkyWater SKY130 process.
Hi Product Hunt, I'm Cascodex, the founder of Cascodex.
If you've designed an analog circuit, you know this feeling: everything passes at typical, then one hot, fast, low-supply corner fails. A single testbench is 45 PVT corners. Checking them, finding out why a spec failed and fixing it without breaking another spec is still mostly done by hand.
So I built Cascodex to automate that loop:
- Upload your SPICE netlists with one small cascodex.yaml listing your specs, and every corner runs in a sandboxed ngspice - When a spec fails, an AI agent traces it to the device causing it and explains why, with every claim linked to a measured value - It proposes a netlist change and re-simulates all 45 corners. A fix only counts if every spec passes at every corner - Corner maps, waveforms, run comparisons, shareable reports, and a CLI and GitHub Action to check every corner on every commit
The one rule I built it around: the AI never estimates a number. Everything comes from the simulator.
On our internal benchmark, the agent found a verified fix in 32 of 33 runs, at a median of about $0.08 and under 2 minutes each. To be honest about where it's weaker: harder designs that fail several specs at once are still a work in progress, and that's what I'm improving next.
You can try it free right now on the open SkyWater SKY130 process. There are ready-made examples (5T OTA, Miller OTA, LDO, bandgap), so you don't need to upload anything. If your team uses commercial process kits or simulators, I'd love to talk about running Cascodex inside your own network.
I'd really value your feedback, especially from analog designers. Tell me where it breaks.
And a question for the designers here: what's the corner that has burned you the most?
No comment highlights available yet. Please check back later!
About Cascodex on Product Hunt
“An AI agent that verifies analog chips at every PVT corner”
Cascodex was submitted on Product Hunt and earned 0 upvotes and 1 comments, placing #421 on the daily leaderboard. Analog circuits rarely fail at typical. They fail at the edges: fast silicon, a hot die, a low supply. Cascodex simulates all 45 PVT corners of your design in ngspice. When a spec fails, an AI agent traces it to the device causing it, edits the netlist and re-runs every corner before calling it fixed. On our benchmark it found verified fixes in 32 of 33 runs, at about $0.08 each. Corner maps, waveforms, shareable reports and CI included. Free on the open SkyWater SKY130 process.
Cascodex was featured in Hardware (11.5k followers), Developer Tools (521.5k followers) and Artificial Intelligence (481.6k followers) on Product Hunt. Together, these topics include over 223.9k products, making this a competitive space to launch in.
Who hunted Cascodex?
Cascodex was hunted by Cascodex. A “hunter” on Product Hunt is the community member who submits a product to the platform — uploading the images, the link, and tagging the makers behind it. Hunters typically write the first comment explaining why a product is worth attention, and their followers are notified the moment they post. Around 79% of featured launches on Product Hunt are self-hunted by their makers, but a well-known hunter still acts as a signal of quality to the rest of the community. See the full all-time top hunters leaderboard to discover who is shaping the Product Hunt ecosystem.
Want to see how Cascodex stacked up against nearby launches in real time? Check out the live launch dashboard for upvote speed charts, proximity comparisons, and more analytics.
Cascodex
Hi Product Hunt, I'm Cascodex, the founder of Cascodex.
If you've designed an analog circuit, you know this feeling: everything passes at typical, then one hot, fast, low-supply corner fails. A single testbench is 45 PVT corners. Checking them, finding out why a spec failed and fixing it without breaking another spec is still mostly done by hand.
So I built Cascodex to automate that loop:
- Upload your SPICE netlists with one small cascodex.yaml listing your specs, and every corner runs in a sandboxed ngspice
- When a spec fails, an AI agent traces it to the device causing it and explains why, with every claim linked to a measured value
- It proposes a netlist change and re-simulates all 45 corners. A fix only counts if every spec passes at every corner
- Corner maps, waveforms, run comparisons, shareable reports, and a CLI and GitHub Action to check every corner on every commit
The one rule I built it around: the AI never estimates a number. Everything comes from the simulator.
On our internal benchmark, the agent found a verified fix in 32 of 33 runs, at a median of about $0.08 and under 2 minutes each. To be honest about where it's weaker: harder designs that fail several specs at once are still a work in progress, and that's what I'm improving next.
You can try it free right now on the open SkyWater SKY130 process. There are ready-made examples (5T OTA, Miller OTA, LDO, bandgap), so you don't need to upload anything. If your team uses commercial process kits or simulators, I'd love to talk about running Cascodex inside your own network.
I'd really value your feedback, especially from analog designers. Tell me where it breaks.
And a question for the designers here: what's the corner that has burned you the most?