# AI can now turn a photo of your board into a schematic. Is your trade secret still a secret?

AI tools can now turn a photo of a board into a schematic. Learn when relying on trade secrets alone gets riskier and when to file a patent.

Written by [Fearn.ai](https://fearn.ai/)

Published Nov 2026 - 5 min read

![AI can now turn a photo of your board into a schematic. Is your trade secret still a secret?](https://www.flux.ai/p/assets/blogs/fearn-ai-reverse-engineering/main.webp)

Founders and hardware teams often run into the same question: our board ships inside an enclosure, so why not just keep the schematic locked down instead of patenting the circuit? For years, that was a reasonable bet, because rebuilding a schematic from a populated board meant hand-tracing copper, X-raying the board, and sometimes stripping it layer by layer. That could take a skilled engineer weeks.

AI tools can now do much of that work. As they spread, relying on trade secrets alone to protect circuit designs gets riskier, and filing patents may become the better strategy.

## What AI reverse-engineering tools can do now

Commercial tools now promise to go straight from a photo to a schematic. [TraceAI](https://trytraceai.com/), for example, says you can upload a picture of a circuit board and get back a [netlist](https://www.flux.ai/p/blog/pcb-netlists-explained). Its stated goal is to reverse engineer a board in minutes rather than weeks. [PCB Tracer](https://hackaday.com/2026/02/23/x-ray-a-pcb-virtually/) lets you trace a board from photos and call an AI model to draft the schematic. [re:trace](https://github.com/ericrihm/retrace), an open-source project, goes from a single board photo to a list of components and the connections between them.

The makers of these tools say results depend on image quality and board complexity, and a photo only shows the outer copper layers. However, for a board’s owner, those limits don’t count for much. On many designs the topology and component choices are visible from the outside, and X-ray or CT imaging can fill in the inner layers.

Going from a recovered schematic to a working copy is getting easier as well. A competitor holding a recovered schematic can use an AI design workflow to produce a board of their own.

## What this means for trade secrets

Trade secret law has never protected against [reverse engineering](https://www.flux.ai/p/blog/pcb-reverse-engineering). The Defend Trade Secrets Act expressly says that reverse engineering and independent derivation are not “improper means” (18 U.S.C. § 1839(6)(B)), and state law generally agrees. A competitor who buys your product, opens it up, and figures out the circuit is free to use what they learn.

Hardware usually ships without any signed IP agreements from the buyer, so there’s often no contract to enforce against a customer who tears down your product. NDAs with your contract manufacturers and suppliers only cover leaks from those companies. Copyright may cover your schematic drawings and layout artwork, but not the circuit itself, so a competitor who builds an equivalent board from a netlist may not infringe.

Trade secrets still work well for what a board doesn’t reveal: manufacturing process details, test procedures, calibration data, and supplier relationships. For more on weighing the two, see Fearn’s guide to [choosing between a patent and a trade secret](https://fearn.ai/resources/patent-or-trade-secret-choosing-the-right-protection-for-your-innovation).

**Key Takeaway:** If your circuit is visible on the board you ship, assume someone’s AI tool can eventually turn it into a schematic.

## If you rely on trade secrets

To qualify as a trade secret, information has to be protected by reasonable measures to keep it secret (18 U.S.C. § 1839(3)). For hardware, that means NDAs with everyone who handles your design files, plus access controls on the repository and the [Gerbers](https://www.flux.ai/p/blog/gerber-files-explained).

Physical measures can raise the cost of a teardown. Potting compound, sealed enclosures, removed part markings, and critical traces routed on [inner layers](https://www.flux.ai/p/blog/pcb-stackup-design-guide) all make photo-based tools less useful. None of them stops a determined competitor, who can X-ray or CT-scan the board, strip the potting, or delayer it. These steps are still worth taking, partly because they’re evidence of reasonable measures if you ever need to enforce a trade secret.

## Why patents look better now

A patent covers your invention no matter how a competitor arrived at it, including through reverse engineering or independent development. That’s the core difference from a trade secret, and it matters more now that a photo and an AI tool can stand in for weeks of tracing.

The traditional downside of patenting is that you have to disclose how the circuit works. That downside shrinks when a competitor could recover the schematic from your product anyway.

The main hurdle for circuit patents is obviousness. Many designs combine well-known building blocks, so examiners often reject broad claims, and the claims that survive can be narrow enough to design around. Patents also take time and money, so it makes sense to focus them on the circuits that matter most.

## When to file

Planning to keep a circuit secret now and patent it later can backfire. In the U.S., selling or publicly using a product that contains an invention starts a one-year clock (35 U.S.C. § 102), and once that year runs out, your own launch can block a patent on the design. Showing a working prototype at a trade show or in a crowdfunding campaign can count as a public disclosure too. Many other jurisdictions, including Europe, offer little or no grace period, so disclosing before you file can rule out patents there entirely.

If a circuit might be worth patenting, file at least a provisional application before the product ships or goes on display. A provisional is quicker to prepare than a full application, and it gives you 12 months to file the nonprovisional.

## How to decide

A useful test is to ask what a competitor would learn by photographing your board. Circuits that are visible on the board and will matter for years are strong patent candidates. Layout tricks, manufacturing know-how, and calibration data can stay secret, backed by real secrecy measures. Most companies will end up doing both.

How long a patent takes to issue matters too. Patents often take two or more years, which is a poor fit for a product generation you expect to replace within a year. A board sold on its own, or one that’s easy to reach inside a consumer device, is easier to copy than one sealed in a potted enclosure. Patent budgets go furthest on the designs that set your product apart.

For a deeper look at the tradeoffs for circuit designs, see [our guide on patents vs. trade secrets for circuit designs](https://www.flux.ai/p/blog/pcb-trade-secrets-vs-patents).

*This article is intended as general information and does not constitute legal advice. Talk to legal counsel before making patent filing or trade secret decisions for a particular product.*

[![Fearn logo](https://www.flux.ai/p/assets/misc/people/fearn-logo.webp)](https://fearn.ai/)

This guide was contributed by [Fearn](https://fearn.ai/), the startup-first patent firm. Built by former Big Law patent experts using modern technology, Fearn helps high-growth startups secure top-tier patent protection at startup speed.

Let’s start building

Build your big idea together with Flux. All you need to do is describe what you want to make, or what problem you want to solve. Flux will work with you to make it a reality.

![Close-up photo of a PCB made with Flux, with an ESP32 module, buttons, and a microSD slot.](https://www.flux.ai/p/assets/blogs/miscellaneous/sidebar-cta-pcb.webp) [Try it Free](https://www.flux.ai/signup)

## Related Content

[

![Line illustration of an older electronic product being refreshed for a new production run](https://www.flux.ai/p/assets/blogs/remanufacturing-legacy-hardware-new-run/main.webp)

### Re-Manufacturing Legacy Hardware: Updating Old Boards and Enclosures for a New Production Run

6 Min Read](https://www.flux.ai/p/blog/remanufacturing-legacy-hardware-new-run)[

![A fountain pen signing a line on a sheet of paper](https://www.flux.ai/p/assets/blogs/fearn-ai-patent-law/main.webp)

### Artificial intelligence and patent law: what PCB designers and manufacturers need to know

11 Min Read](https://www.flux.ai/p/blog/ai-patent-law-pcb-designers)[

![Line illustration of design files moving from an old dusty boxy computer monitor along a glowing path into a modern clean 3D view of a circuit board seated inside its enclosure](https://www.flux.ai/p/assets/blogs/move-legacy-hardware-off-discontinued-tools/main.webp)

### Moving Legacy Hardware Designs, Boards and Enclosures, Off Discontinued Tools

5 Min Read](https://www.flux.ai/p/blog/move-legacy-hardware-off-discontinued-tools)[

![Line illustration of a circuit board inside an open enclosure where one old chip is lifted out and a new replacement chip lowers into place, with dashed fit lines showing the board still matches the enclosure mounting posts](https://www.flux.ai/p/assets/blogs/revive-product-end-of-life-parts-enclosure-fit/main.webp)

### Reviving a Product With End-of-Life Parts: Redesigning the Board Without Breaking the Enclosure Fit

5 Min Read](https://www.flux.ai/p/blog/revive-product-end-of-life-parts-enclosure-fit)[

![Line illustration of a small round smart home door sensor in exploded view](https://www.flux.ai/p/assets/blogs/first-founder-smart-home-sensor-board-case/main.webp)

### From Idea to Pre-Orders: How a First-Time Founder Designed a Smart Home Sensor, Board and Case

5 Min Read](https://www.flux.ai/p/blog/first-founder-smart-home-sensor-board-case)[

![Line illustration of a laptop showing code brackets connected by a glowing line to an exploded view of a small handheld device](https://www.flux.ai/p/assets/blogs/software-engineer-first-hardware-product/main.webp)

### A Software Engineer's Guide to Designing a First Hardware Product, From Circuit to Enclosure

5 Min Read](https://www.flux.ai/p/blog/software-engineer-first-hardware-product)

See All (6)

## Let’s start building.

Let’s build your idea together. All you need to do is describe what you want to make, or what problem you want to solve. Flux will work with you to make it a reality.

[Try it Free](https://www.flux.ai/signup)

 ![Close-up of a Flux-designed printed circuit board](https://www.flux.ai/p/assets/homepage-v2/mission-visual.webp)