Protect Authors. Prove Process.
You wrote it, and someone is about to ask you to prove it. I built WritersLogic so the answer is a record instead of an argument: tamper-evident proof of your own effort, and an honest way to say where AI helped.You wrote it, and sooner or later someone is going to ask you to prove it: a professor, a client, an editor, a platform. I built WritersLogic so that the answer is a record instead of an argument. It gives you tamper-evident proof of your own effort, and it gives you an honest, checkable way to say where AI helped and where it did not. Not detection. Not accusations. Evidence you can hand over.
David Condrey, Founder
How This Started
In August 2025, Medium’s CEO published a post asking writers how AI should be used on the platform. I wrote a long comment. Three people replied to say it must have been written by AI, one of them citing a detector reading of 91%, and not one of them answered a single thing I had argued.In August 2025, Medium’s CEO published a post asking writers how AI should be used on the platform. I wrote a long comment arguing that the line that matters is not human versus machine, it is whether a person did the work. Three people replied. “Honestly, this sounds like it was written by AI.” “Oh, it was. Without a doubt.” “91% AI according to” a detector. Not one of them answered a single thing I had argued.
I do use AI. As an assistant that helps me write, not as a replacement for writing. That distinction was the whole point of the comment, and it was the one thing nobody would engage with, because there was nothing I could point at to settle it.I do use AI. As an assistant that helps me write, not as a replacement for writing. That distinction was the whole point of the comment, and it was the one thing nobody would engage with. There was no way to answer the accusation, because there was nothing to point at. On a platform whose own distribution rules require that a story “does not appear to be generated by” AI, appearing is enough. The accusation costs nothing to make and there is no defense against it.
I went looking for something that would settle it either way. There wasn’t anything. So I stopped writing and built it.
"Was this written by AI?" is the wrong question. Writing with AI isn't cheating. Claiming AI-authored work as your own is. The distinction that matters is between AI-assisted and AI-authored, and that distinction requires evidence, not guesswork."Was this written by AI?" is the wrong question, because it assumes a binary that no longer exists. Writers brainstorm with AI. They use it to break through blocks, to rephrase awkward sentences, to explore structures they hadn't considered. That isn't cheating. What matters is whether the writer shaped the final work, and whether they're honest about the process. The distinction between AI-assisted and AI-authored is everything, and that distinction requires evidence, not guesswork.
So I asked a different question: what if you could say how you worked, and the tool could back you up with the record? Every time you paste, WritersProof asks where it came from. How you type, every pause and every revision are measured as you go, never what you type. If you paste a paragraph from AI and then go back and rewrite most of it, the record shows exactly that: AI-assisted, not AI-authored. Your own writing either supports what you declared or it does not.
That is what WritersProof does. It does not judge and it does not detect. It records how your document was actually made, your natural writing rhythm, what you pasted and what you said about it, how deeply you revised, and seals that record so it cannot be altered afterward. The proof travels with your document. Anyone can check it. You stay in control of your own story, and the evidence keeps it honest.
What This Means for You
The old question, human or machine, is broken. The new standard is showing your work.
You paste a paragraph from ChatGPT, rewrite half of it, restructure the argument and fold it into your own narrative. That is real creative work. Today's tools cannot tell it apart from paste-and-submit. WritersProof can, because it records how much you revised after every paste.
Each time you paste, WritersProof asks where it came from: your own draft, a quote, AI text, or something else. How you write afterward either backs that up or contradicts it. Honesty earns you stronger evidence.
A detector gives you a percentage and nothing else. No method, no explanation, no way to answer it. WritersProof gives you a detailed record: what you typed, what you pasted, what you revised and how deeply. Evidence you can hand over, not a score you have to argue with.
As a student, you can use AI to learn and still prove you did the work. As a freelancer, you can show a client how the piece was made. As an institution, you can set clear, evidence-based policy instead of trusting a detector. Everyone at the table sees the same record.
You deserve protection, and honest AI use deserves a real way to show itself. WritersLogic gives you both: tamper-evident proof that defends the work you did, and a clear record of where AI fit into it.
What You Can Count On
Not abstract values. Concrete commitments you can check.
Proof you can examine, not scores you must trust
We do not detect AI. We prove you wrote it. Our methods are published, and every report shows exactly which evidence backs your proof.We do not detect AI. We prove you wrote it. Our methods are published and open to audit. Every report shows exactly which evidence backs your proof: which writing signals were recorded, how confident each one is, and where the evidence is thin. If we cannot substantiate a claim, we do not make it.
In practice: WritersProof gives you an Authorship Report that breaks down exactly which signals support your proof: typing patterns, revision depth, session timing, and more. The protocol is open source under Apache-2.0, so anyone can check how that report was produced.
Your writing stays on your machine
WritersProof runs on your own computer. Your writing rhythm, your pauses and your revisions are recorded locally on your device. Your text stays on your machine; recording and signing run locally; evidence is shared only when you choose to back it up or share it. That evidence is a small, tamper-evident record that the work happened, never a copy of the work itself.
In practice: Watch your network traffic while you write in WritersProof. You will see small signed proof records going out, and never your text. We cannot read what you write, because the software is built so that your text never reaches us, even if someone compelled us to try.
If the science doesn't support it, we don't claim it
WritersProof rests on tamper-evident evidence of your effort, not statistical guesswork. It records the full story of your writing: how you typed, what you pasted, what you revised and how deeply. You declare how you worked, and the evidence either supports you or it does not. When something cannot be substantiated, we do not report it as proven.
In practice: A WritersProof Authorship Report never says "87% human." It shows which passages you typed, which you pasted and what you said about their origin, how heavily each section was revised, and whether your writing patterns back up your declaration. Evidence you can hand over, not a score you have to trust.
Your proof needs open standards, not a vendor you depend on
Proving who wrote something is too important to be controlled by any single company, including ours. That is why we published the protocol behind WritersProof as an IETF Internet-Draft, why the code is fully open source, and why your proof is built to travel with the content-authenticity standard that platforms and publishers are adopting. If WritersLogic disappeared tomorrow, every proof you have created would still verify.
In practice: Every proof you create is timestamped on a public ledger and signed with the same security technology that protects your bank and your email. Anyone can verify it with the open-source reference implementation: no account, no subscription, no dependence on us.
How WritersProof Works
Real proof of your effort, not statistical guesswork
The WritersProof Engine
Your Natural Writing Rhythm
The measurable effort that only a real person produces while writing
- •Your unique typing speed and rhythm between keys
- •Thinking pauses, where you stop to consider, not just where you stop typing
- •Revision sequences that reflect live decision-making
- •A writing fingerprint unique to each session and writer
Multi-Layered Security
Four independent signals sealed together. Tamper with any one, and the whole proof breaks
- •Document content fingerprint
- •Your writing rhythm signature
- •Author identity verification
- •A public timestamp anyone can check
Tamper-Evident Timeline
A chronological record that proves the order and timing of your writing
- •Each writing event links to the previous, forming an unbreakable chain
- •Impossible to fake after the fact
- •Any insertion, deletion, or reordering invalidates the proof
- •Chain anchored to a public ledger, so anyone can verify it without asking us
Tamper-Evident Proof
Tamper-evident evidence of your process and effort
- •Public timestamps that prove when your text existed
- •Verified proof linking your text to your identity
- •Immutable audit trail showing the complete writing process
- •Travels with the content-authenticity standard platforms and publishers are adopting
What You Get
Desktop apps and an open protocol, each built for a different need, all resting on a real record of your writing
WritersProof
A desktop app that records your whole writing process as evidence you can hand over. Your typing rhythm, what you pasted and where it came from, your revisions and your thinking pauses are sealed into a tamper-evident Authorship Report that shows AI-assisted work for what it is, and AI-authored work for what it is.
Everything stays on your machine. Paste something from AI? The app asks where it came from. Rewrite it heavily? The record shows exactly that. An honest process, and proof you can show.
Learn about WritersProofWritersProof Protocol
The open protocol under WritersProof. It defines how your writing evidence is recorded, sealed against tampering, and timestamped in public so anyone can verify it without asking us. Built on proven security technology, purpose-built for proving authorship.
Published as an IETF Internet-Draft and built to travel with the content-authenticity standard major platforms use. Apache-2.0 licensed, because proof of your own work should never be someone else's property. If WritersLogic disappeared tomorrow, your proofs would still verify.
View on GitHubChronograph
A research-grade writing analysis tool built for academics. Hardware-secured data, precise timing measurements, built-in privacy protections, and AI-powered cognitive analysis, all without your content ever leaving your machine.
Built for studies in cognitive science, writing process research and writing pattern analysis. Every data point is signed by the hardware that recorded it, so a forged record cannot pass as real.
Learn about ChronographWritersEye
A desktop writing app where AI analyzes and never generates. It recognizes your voice, flags where a draft drifts from it, and produces a detailed authorship report, so you, your editor or your publisher can see how the writing holds together.
One-time purchase. No subscription. AI helps with analysis, spotting style drift, flagging inconsistencies, assembling the report, and never writes a single word of your content. Your voice stays yours.
Learn about WritersEyeProof That Outlasts Us
Your proof stays valid whatever happens to any single company, including this one
IETF Internet-Draft
The protocol behind WritersProof, published as an individual IETF Internet-Draft (not an IETF standard). It specifies an open format for tamper-evident evidence of the writing process that any implementation can follow, so your proof never depends on our software to be checked.
View on IETF DatatrackerC2PA Compatible
Built to work with the Coalition for Content Provenance and Authenticity (C2PA) standard that platforms and publishers are adopting. As that ecosystem grows, your WritersLogic proof plugs straight into it. Proof of your authorship travels with your content.
Open Source Core
The WritersProof protocol and reference implementation are Apache-2.0 licensed, and the analysis method is published. You can audit every line of code that touches your writing. Trust in a tool that vouches for you has to be earned, not demanded.
View on GitHubPublished Research
Detailed writeups on the protocol, the security behind WritersProof, and the writing pattern model. Not behind a paywall, not in marketing-speak. Real documentation you can evaluate for yourself.
Read the researchWho's Behind This
David L. Condrey
Founder and Principal Engineer
I've spent my career at the intersection of software engineering, security, and the written word, and I brought that rigor to authorship verification.I've spent my career at the intersection of software engineering, security, and the written word. Before WritersLogic, I worked on systems where getting verification wrong had real consequences, and I brought that same rigor to proving authorship.
I started this project because I saw writing collapse into a binary: human or machine. But writers have always collaborated. AI is one more collaborator. What was missing was infrastructure to make the human effort visible.I started this project because I watched writing collapse into a binary: human or machine. But that binary never reflected how writing actually works. Writers have always collaborated, with editors, workshops, beta readers. AI is one more collaborator. The people AI frees are writers who have lived through real experience but lacked the craft to compete with polished imitators. What was missing was infrastructure to make the human effort visible.
So I drafted the protocol, open-sourced it, and built a desktop app that records how you actually write, keystroke timing, paste declarations, revision depth, and seals it into tamper-evident proof.So I drafted the protocol, published it as an IETF Internet-Draft, open-sourced the implementation, and built a desktop app around it that records the full story of how you write, keystroke timing (never what you type), paste events with origin declarations, revision depth, and seals it into tamper-evident proof anyone can verify. Not to catch anyone. To make your effort visible.
Get in Touch
Questions about how it works? Want to talk about the science, the standards, or how to bring proof of authorship into your platform? I would like to hear from you.