Common questions for developers

No. Patent PreCheck is an informational tool — not a law firm and not a substitute for a licensed patent attorney. Scores and reports help you prepare; an attorney should review anything you plan to file.

Software can be patented when it meets USPTO requirements: patentable subject matter (not a bare abstract idea), novelty (§102), non-obviousness (§103), utility, and adequate disclosure (§112). Use our free analyzer or read the full guide at /can-my-code-be-patented.

AI-assisted code can still support a patent when a human inventor conceived the claimed invention and can document specific technical choices, rejections of AI suggestions, and measured improvements. The USPTO requires named human inventors — AI cannot be listed as an inventor.

Alice Corp. v. CLS Bank (2014) established that abstract ideas implemented on generic computers are not patentable. Software must show a specific technical improvement or unconventional computer implementation — not routine automation of a mental process.

Your code is analyzed in-session and not stored on our servers after the result. We hash inputs for abuse prevention but do not retain raw source after the free score completes.

Your invention text is saved securely for the 30-day coaching window so you can resume editing. It is used only for your review and is not used to train third-party models. See our Privacy Policy.

The USPTO does not accept third-party score reports as filing documents. Your report is a preparation tool — it helps you strengthen claims and documentation before you or your attorney file.

Scores reflect comparison against our continuously updated prior-art index — USPTO filings, PTAB decisions, abandoned applications, and more — plus USPTO patentability pillars. They are directional, not guarantees. Read /methodology for details.

No. We do not use your uploaded code or invention text to train models. Analysis runs against our indexed prior art corpus — not your content as training data.

Live pillar-by-pillar coaching on your code, human vs. AI contribution analysis, 200-match prior art deep search plus extended report (up to 500 refs), score updates as you edit, Word filing draft, tracked-changes export, ZIP package, and 30 days to return and re-edit.

Prior art is any public disclosure that existed before your filing date: issued patents, published applications, papers, products, and open-source repos. If your invention was already disclosed, it may lack novelty (§102) or be obvious (§103) in view of that art.

Yes. Public open-source repositories, README files, and released packages can be prior art if they disclose the same or similar technical features before your priority date.

A public GitHub repo can be cited as prior art if it predates your filing and describes the same invention. Private repos that were never disclosed generally are not public prior art — but confidentiality matters before filing.

35 U.S.C. §101 asks whether the invention is patentable subject matter. Pure abstract ideas, mental processes, and generic computer implementations of them are often rejected — especially for software after Alice.

§102 requires that no single prior art reference fully discloses every element of your claim before your filing date. Even one matching reference can anticipate your invention.

§103 asks whether a skilled person would combine prior art references to arrive at your invention. Obvious combinations of known techniques are not patentable even if no single reference discloses everything.

§112 requires your specification to teach someone skilled in the art how to make and use the invention without undue experimentation. Thin or purely functional descriptions weaken filing readiness.

Patent PreCheck's inventorship score estimates how well your disclosure documents human-directed technical decisions vs. AI-generated boilerplate. It is guidance for preparation — not a legal finding.

A generic UI component is often abstract or obvious. A component tied to a specific technical problem (latency, security, data integrity) with documented human design choices may fare better. Run the free score on the underlying algorithm, not just JSX.

Mathematical algorithms alone are usually abstract. An algorithm applied as a specific technical solution with measurable improvement (speed, accuracy, resource use) may be eligible — if it is also novel and non-obvious over prior art.

Training pipelines and inference architectures can be patented when they solve a technical problem in a non-obvious way. Generic "apply ML to X" claims often fail Alice and §103. Document what is structurally different about your approach.

No — but public launch, blog posts, and App Store listings can become prior art against yourself if you file later. File a provisional or complete a prior art search before broad public disclosure if IP protection matters.

A provisional secures a priority date for 12 months with a lower filing fee and informal format. It is not examined and expires unless you file a non-provisional. Patent PreCheck helps you prepare stronger disclosure before that step.

Typical non-provisional software applications run roughly $8,000–$17,000 in attorney fees plus USPTO fees. Patent PreCheck ($0–$69.95) helps you arrive prepared so those hours focus on claims strategy, not basic education.

Google Patents is a free search database. Patent PreCheck scores your actual code against USPTO pillars, estimates inventorship patterns, coaches improvements, and exports filing-ready documents.

Evalify analyzes pitch decks for investor FTO at ~$249/report. Patent PreCheck analyzes source code with interactive coaching and human vs. AI inventorship scoring at $69.95 one-time.

Patentable offers free prior art search and $299 provisional generation. Patent PreCheck adds USPTO pillar coaching on your code, inventorship analysis, and tracked Word exports at $69.95.

PowerPatent is an AI drafting workflow (~$199/invention or pro subscriptions) built around disclosure questionnaires and draft documents. Patent PreCheck scores and coaches your source code first — USPTO pillars, inventorship, and prior-art overlap — so you know whether drafting spend is worth it.

PQAI is freemium prior-art search over patents and papers. Patent PreCheck analyzes your source code with USPTO pillar scores, inventorship signals, and coached edits — not search-only results.

Patent when competitors can reverse-engineer the product and you are willing to disclose the mechanism. Prefer trade secret when the value is confidential know-how that would teach rivals if published. Run a free patentability score before deciding either way — see /blog/patent-vs-trade-secret-code.

No. You control distribution. Share outputs with any attorney you choose. A vetted attorney directory is on our /roadmap.

Review our public sample report at /sample-report, the self-scored /case-study, /methodology, and /security. Firms evaluating tools for internal committees do not need to upload client materials — clients run the tool on their own code and share deliverables with counsel when ready. See /attorneys#for-counsel.

No. We estimate how well a disclosure documents human-directed technical decisions vs. AI-assisted boilerplate. That is preparation guidance for you and your attorney — not a legal inventorship finding. Final inventorship analysis belongs with licensed counsel before filing.

The directory is on our roadmap. When it launches, inventors who complete Interactive Code Review can choose to contact listed firms — we will not forward reports without the client’s direction. Listings will highlight software/AI experience and intake expectations. Attorneys can apply at /attorneys.

Use Patent PreCheck first when you want a fast, code-based read on patentability pillars, prior art overlap, and documentation gaps — before spending on prosecution. Hire counsel for claim drafting, inventorship/legal judgment, filing strategy, and representation. Patent PreCheck helps you arrive prepared; it does not replace an attorney.

Run the free score, note weak pillars and closest prior art, then use Interactive Code Review if you want coached edits and a filing-ready package. Bring the report, code pack, and inventorship notes to your attorney so the first call focuses on strategy — not re-explaining your build from scratch.

Scores are directional, not pass/fail. Rough bands: 70+ suggests strong prep potential; 50–69 often means promising ideas with specific pillars to strengthen; below 50 may indicate serious prior-art or eligibility headwinds worth addressing before filing fees. Read /methodology for how scores are calibrated.

Each free score gets a report ID (e.g. PPC-2026-05-29-ABCDE). Use Save report by email on /analyze, then reopen it anytime at /retrieve with that ID — or use the View report online link in your email.

Yes — within 30 days of payment you can resume coaching, edit source code, and re-finalize to regenerate your filing package.

Any — Cursor, GitHub Copilot, Claude, ChatGPT, Codeium, Windsurf, etc. Paste the resulting source code and declare how AI was used.

Paste source text directly. Common languages include JavaScript, TypeScript, Python, Java, Go, Rust, C++, and more. The analyzer looks at technical content, not file extension alone.

Most free scores complete in under 60 seconds depending on code length and prior art search depth.

The free tier shows a prior-art count plus a teaser of up to 20 neighbor titles. Paid Interactive Code Review includes a 200-match deep search with overlap narrative, plus an extended deep report (up to 500 references) in the background.

Patentability reflects novelty, non-obviousness, and eligibility strength. Filing readiness (§112) reflects whether your disclosure is specific enough to support enforceable claims.

FTO asks whether shipping your product might infringe someone else's existing patents — different from whether your invention is patentable. Evalify focuses on FTO; Patent PreCheck focuses on patentability prep.

Only the first inventor to file (with exceptions) generally gets the patent for the same invention. Document conception dates and contributions if you collaborate.

Copyright protects expression of code automatically; patents protect functional inventions. They complement each other. Software startups often use both for core algorithms.

No. In Thaler v. Vidal (Fed. Cir. 2022), only natural persons qualify as inventors. AI-assisted code is fine — but you must document human conception. See /legal-intelligence for daily updates.

If you opt in on the free score or purchase a review, we store email to deliver reports and session links per our Privacy Policy. We do not sell email lists.

Yes. You can run a free patentability check right inside the AI coding assistant you already use — Claude, Cursor, ChatGPT (Codex), and others. You add one line once, then just ask your assistant to “run a patent precheck on this.” No account and no setup keys. Step-by-step on /integrations.

Yes. The same scoring engine is available as a free command-line tool (CLI), an MCP server your AI assistant can call, and a web API. See /integrations for the one-line install and POST /.netlify/functions/analyze usage.

Yes. You can add a check that runs by itself every time you update your project (developers call this “CI”). Grab the ready-made setup from /integrations — no need to hand us your private code.

Scoring and prior art signals emphasize US law and USPTO practice. Many principles overlap internationally but filing rules differ by jurisdiction.

A low score is useful — it suggests weak pillars or prior art overlap before you spend on filing. The Interactive Code Review coaches specific edits to raise scores and documentation quality.

Yes. Use /contact for team plans, portfolio tracking, and team inventorship features on the roadmap.

Daily. We ingest Tier 1 and Tier 2 patent law feeds (USPTO, Federal Circuit, IP blogs) described on /legal-intelligence.

ZIP package with README, score comparison, filing-ready Word draft, tracked-changes Word doc, branded PDF scorecard download, and email delivery.

Yes. After you analyze on /analyze, use Download PDF scorecard on the results page. You can also email yourself the report for a shareable online link at /r/{id}.

Interactive Code Review is one-time $69.95. Legal-change monitoring is included for 30 days after your paid review (and on each Portfolio invention while subscribed). Continue alerts for $19.95/month per invention from your scorecard.

Patent PreCheck was founded by a team of experienced entrepreneurs and software engineers. See /about for our mission, methodology, and contact.