Straight answers

Confidentiality

Where does my client's file history go?

If you run a local model through Ollama, it stays on your machine. The scanned pages are recognised by an OCR engine on the same computer, the reading runs on a model on the same computer, and the result is stored with the matter on your disk. There is no server of ours anywhere in that path, because there is no server of ours at all for this work.

If you configure a remote provider with your own key, the recognised text of the papers it reads 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?

Local processing is the conservative choice for the file history of anything unpublished, though it is not a compliance certificate: the firm's own endpoint security, backups and client instructions still apply. For reading, it gave up little in the published measurement: a 9-billion-parameter model on a laptop found every rejection in two real file histories, and missed one reference that two remote models found. Whether a particular remote provider is acceptable for a particular matter is your professional judgment to exercise, and it depends on facts we do not have.

The measurement is small and is published with its method on the evidence page.

Does the tool phone home?

Not for reading or 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 file history, 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.

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 file history and no specification ever reaches this site: it renders pages and holds nothing else. The program you download makes no analytics call of any kind, and the one exception it does make is the licence renewal described above.

What it reads, and what it refuses to do

What does it do with an Office action?

It reads it and reports what it says. Attach a Patent Center "download all documents" PDF and the tool lists every paper in the file, picks out the Office actions, the responses and the Notice of Allowance, recognises the text of their scanned pages, and reports each rejection: the claims, the statutory ground, the references applied and the examiner's sentence it came from, with each quotation checked against the paper and marked verified or not.

Where the application was allowed it also lays out the rounds and marks every limitation of the allowed claims as filed with the application or added during prosecution, and quotes the examiner's stated reasons for allowance. One real reading is printed unedited.

Does it draft a response to an Office action?

No. It does not draft a reply, an amendment or remarks, and it does not suggest how to answer a rejection. The USPTO's rules list drafting a reply to an Office action that may require written argument as practice before the Office (37 CFR 11.5(b)(1)), and the reply is signed by the practitioner who stands behind it. The tool gives that practitioner the record, laid out and quoted.

Does it say whether a rejection was right?

No. It reports what each paper says and never whether a rejection was correct, an amendment necessary, or an argument likely to succeed. What the record means for a matter is the practitioner's reading.

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.

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. On the reading side it reports what the papers say; on the drafting side, described at patentcontinuation.com, it scores drafting quality against the specification you supply.

Does it track deadlines?

No. It shows the date of each paper as Patent Center records it, and it does not compute reply periods, extensions or deadlines. It is not a docketing system, and a deadline should come from yours.

Is the output filing-ready?

No. It is R&D-grade and every output says so. A reading is a laid-out, quoted copy of what the papers say, for a licensed practitioner who checks it against the papers and exercises their own judgment. Where a quotation is marked not verified, look at the paper.

It does not replace an attorney

It is a reading and 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.

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 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.

How do I give it a file history?

In Patent Center, download the application's documents with "download all documents". In the app, open a matter and attach that PDF with Attach file history. Reading a file history is in the app's interface; the command line does not do it.

The pages are scanned images, so reading them needs an OCR engine on your machine: tesseract, if installed, or a PDF you have already run text recognition on. If neither is available, the app says so rather than reporting an empty file.

The program is called Continuation Drafter. Is that the right download?

Yes. Reading a file history is one of the things Continuation Drafter does, and the same app drafts continuation claims, which is where its name comes from. This site describes the reading; patentcontinuation.com describes the drafting.

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.

What hardware do I need?

For reading, the published measurement ran a 9-billion-parameter model on a laptop. For local models in general, memory is the binding constraint, and 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 reading take?

Minutes rather than seconds for a typical file, most of it spent recognising scanned pages. We have not measured it systematically, so we give no figure for it, and we do not claim it saves you time, which is a different claim we have not measured either.

The project

Is this open source?

No. The program 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.

Why are there two websites?

Because practitioners arrive with two different questions. This site is for the one who has an Office action and wants the record laid out; patentcontinuation.com is for the one drafting a continuation. It is one program, and both sites are published by Obviously Not, Inc.

Why should I believe the evidence?

Because you can check it. Both file histories are public and named, the method is published, the runs that did worse are in the same table as the ones that did well, and the limits are stated on the page rather than in a footnote: two files is a first check, not a survey.