Straight answers

Confidentiality

Where does my client's specification go?

If you run a local model through Ollama, it stays on your machine. The specification is read from your disk, passed to a model running on the same computer, and the drafted claims are written back to your disk. There is no server of ours anywhere in that path, because there is no server of ours at all for drafting.

If you configure a remote provider with your own key, the specification goes to that provider, exactly as it would if you pasted it into their web interface. That is a different arrangement and you should treat it as one.

So which should I use?

The local path is the one to use for anything unpublished or privileged. The remote path is faster and the stronger models draft better, which is a real tradeoff and not a trick question. Whether a particular remote provider is acceptable for a particular matter is your professional judgment to exercise, not ours to pre-empt, and it depends on facts we do not have: the client, the engagement terms, and the provider's own contractual commitments.

Does the tool phone home?

Not for drafting, ever. There is one exception and it exists only if you buy a licence: the binary renews that licence when it is close to expiring, roughly once a month, on launch. It sends the licence key and nothing else. No specification text, no matter identifiers, no usage data. Setting CD_NO_LICENSE_RENEWAL=1 switches it off. Without a paid licence, the binary makes no such call at all.

Since v0.4.0 you can also ask it to check for a newer version, from Settings or with continuation-drafter update. That call happens only when you ask, never on launch or on a timer. It is one request to GitHub's public releases page carrying nothing about you, and offline mode refuses it. An update is installed only after its signature and checksum verify, so a tampered release is refused rather than installed.

For a firm's IT or security review, every connection the program makes, where it keeps its files and how a download is signed are on one printable page.

Is there any analytics on this website?

Yes, and it is worth being precise about what that means. This site uses Cloudflare's web analytics: page views, page timings, and coarse data such as country and browser. It is cookieless, it does not fingerprint your device, and it builds no profile across visits or across sites. A cookieless one was chosen deliberately, because a practitioner's browsing is itself suggestive of client work.

The website and the tool are separate, and only the website is measured. No specification and no draft ever reaches this site: it renders pages and holds nothing else. The binary you download makes no analytics call of any kind, and the one exception it does make is the licence renewal described above.

What it does, and what it refuses to do

It does not assess patentability, novelty or non-obviousness

It is built not to. It does not assess patentability, novelty or non-obviousness, and it does not search prior art. Those are legal determinations. This tool drafts claim language and scores drafting quality against the specification you supply: whether a claim's elements and their particular combination are described there, whether it is definite, and how it differs from the claims already filed. The scoring prompt refuses patentability questions on purpose, and the refusal is printed in the output rather than hidden.

The authorities those determinations actually rest on are published separately and free of the tool: twenty-eight decisions and rules, each quoted from its primary source and checked against a cached copy by script rather than by eye. Reading them is how you make the judgment yourself. Nothing there assesses anyone's claims either.

The output is not filing-ready

It is R&D-grade and every output says so. It is drafting input for a licensed practitioner who reviews every limitation against the specification and exercises their own judgment. Treat it as a first pass from a capable but unsupervised junior: worth having, never sent out unread.

It does not replace an attorney

It is a drafting aid for licensed patent attorneys and agents, and for people working under their supervision. It is not addressed to the general public and it does not practise law.

What is it actually good at?

Producing a broad set of continuation claim candidates from a specification you have already filed, quickly, so you are editing rather than starting from a blank page. It also flags per-claim defects (antecedent basis, support, indefiniteness) for you to resolve. Our own findings say the critique is the load-bearing part and the numeric score is the weakest part, and the benchmark page says so too.

If what you want is the doctrine rather than the tool, that is a different thing and it is also free: the continuation course sequences those same authorities, one question answered before anything else and then five steps in the order a family is built. It sells nothing and requires no account.

Can it read the prosecution history?

Yes, since version 0.3.0. Attach the parent's Patent Center "download all documents" PDF. Patent Center's pages are scanned images, so the tool reads them with an OCR engine on your own machine (tesseract, if installed, or a PDF you have already run text recognition on), and no page leaves the machine for that step. It reports what each paper says: the rejections, the references each relies on, the examiner's reasons for allowance, and whether each limitation of the allowed claims was filed with the application or added during prosecution. It does not say whether an amendment was necessary or a rejection correct; that is the practitioner's reading. How well it reads an Office Action was measured on real files.

Does it search for prior art, or compare my claims against the references?

No to both. It reports what the file's own papers say about the references in them, and it never names a reference the file does not contain. It does not search, and it does not compare any claim against any reference: that comparison is the practitioner's.

Running it

Do I need to use the Terminal?

On a Mac, no. Download the app from the download page, open it, drag Continuation Drafter to your Applications folder, and double-click it. The drafting interface opens in your browser. macOS asks once whether you are sure you want to open something downloaded from the internet; the button is Open.

On Windows and Linux there is no app bundle, so you start it once from a terminal, or on Windows by double-clicking the .exe, which opens a console window. One command, ./continuation-drafter serve, and the same interface opens in your browser.

Everything the command line does is in that interface. Command-line binaries are published for macOS, Linux and Windows if you would rather work at a prompt, and they are the right choice for a script or CI.

The app has no Dock icon, because it runs no window of its own: the interface is the browser tab it opens. Press Stop in that tab to shut it down.

What does it cost?

The tool is free to download and use. If you run local models, there is no marginal cost at all. If you use a remote provider, you pay that provider directly with your own key; nothing routes through us and we do not mark it up.

Either way, every run ends by reporting what it used and what it cost. On your own machine that report sits beside what the same work would cost on a leading hosted model at its list price; through a provider it is the cost that provider reported. The web interface also keeps a running total for each matter.

What hardware do I need?

For local models, memory is the binding constraint, and the honest answer is that it depends on your specification's length rather than its complexity. The setup page asks what machine you have and names a model that fits, with the measured tradeoff stated rather than hidden.

How long does a draft take?

Across 131 measured runs over 10 real patent families and 15 remote models, the median was 78 seconds to a full set of about 13 claims, and 84% of runs finished inside three minutes. Two caveats that matter: those runs used remote models, so local inference on your own machine will be slower and depends on your hardware; and this measures how long the tool takes, not how much time it saves you, which is a different claim that we have not measured and therefore do not make.

Since v0.5.0 the app splits a run at the point you decide: the first press proposes the directions and drafts nothing, and the drafting time is spent only on the directions you keep.

My specification is enormous. Is that a problem?

Less than you would expect. Across those same runs, specification length did not predict drafting time (the correlation was -0.118 over specifications from about 17 pages to about 240). The longest specification tested drafted faster than a third of the smaller ones. Which model you choose matters far more than how long your document is, though the model's context window does have to fit the specification, which is what the setup page is for.

I have the claims. How do I get them into Word?

continuation-drafter export --draft claims.txt --out claims.docx, or pipe a drafting run straight into it. Your claim numbering is preserved exactly as drafted and never renumbered, since a continuation's claims are commonly numbered on from the parent. LaTeX is supported too, with --out claims.tex.

The exported document carries the maturity notice inside it, deliberately: an exported file gets read by people who never saw the terminal that produced it.

The project

Is this open source?

No. The binary is free to download and use, and it is proprietary. The source is not published.

Why is it free?

It is free because it is research and development: R&D-grade, and the point now is for practitioners to use it and say what is wrong with it. Nothing you do in it is collected, so it is not paid for with your data. A paid tier is possible in the future, but there are no plans for one at this time.

Who is behind it?

It came out of internal R&D on continuation drafting and was extracted into a tool once it drafted end to end. The benchmark that the benchmark page reports is the same one used to decide which models ship in the guided setup.

Why should I believe the benchmark?

Because you can check it. The method is published, the models and families are named, and the limits are stated on the page rather than in a footnote: a single AI judge cannot reliably rank near-equal drafters, so the scores separate gross differences and nothing finer. That is a weaker claim than the ones you will read elsewhere in this category. It is also one you can verify.