🔍 Read the full analysis: What It Took To Make My AI-Coded PDF Editor Dependable on ThorstenMeyerAI.com
Get business pricing on monitors, keyboards and dev gear
- Business-only prices and quantity discounts
- Tax-exempt purchasing
- Multiple users, one account, clear invoices
TL;DR
Thorsten Meyer reports that six days of testing and revisions turned Docestra from a one-day AI-built PDF editor into a more thoroughly tested private tool. Tests across real PDFs uncovered save, encryption and redaction problems; the account also says the developer has not measured time or cost savings.
Thorsten Meyer says six days of testing and revisions turned Docestra, his AI-coded PDF editor, from a working first version into a more dependable private tool. In an account published November 13, he describes problems uncovered by tests of real documents, including a save bug, limits on encrypted files and flaws in an early redaction design.
Meyer says he gave an AI coding agent a one-sentence brief on November 6: to build an Acrobat-like editor for his own use. The first version was committed at 9:47 p.m. Central European Time that day. It could open PDFs, add annotations and signatures, fill forms, rearrange pages and save files. Version 1.5.1 followed on November 12.
A robustness run involving 930 real PDFs found that saving some files could produce a broken cross-reference table, causing markups to disappear. Meyer says he changed the save process to avoid compressed object streams. Testing also found that 52 of 69 encrypted PDFs did not permit annotations under their authors’ restrictions. Because the PDF library could not re-encrypt edited files, Docestra was changed to require confirmation before unlocking such a file and to save an unprotected copy without overwriting the original.
The first version’s white-out feature only covered text; it did not remove the underlying content. Meyer says an independent AI review identified 14 problems in the first review round and 13 in the second for the initial true-redaction design. He then replaced it with an approach that converts affected pages to images and adds a searchable text layer. In a test across 917 PDFs, Meyer reports that this design produced zero detected leaks. He also says the project grew from 17 automated unit tests in its first version to 745 in version 1.5.1.
Docestra · Nov 6–13, 2026
From one prompt to a tool I can rely on
What the first version gave me, what real files broke, and what it took to fix it, from my project log.
What the tool rests on (schematic)
Built bottom up. The top layer is a plan, not a finished feature.
What happened, by date
- Nov 6One-sentence brief; first working version committed the same evening
- Nov 6Test over 930 real PDFs finds a save bug that dropped markups
- Nov 752 of 69 encrypted PDFs forbid annotations: unlock only into a new copy
- Nov 8Real redaction: review finds 14, then 13 problems; simpler design, 0 leaks in 917 PDFs
- Nov 12Version 1.5.1: 745 unit tests, 32 browser test suites
- Nov 13Version 2.0 starts: one shared PDF engine for our AI agents (in development)
Source: Docestra project records (development log, plans for versions 1 and 2.0, commit history), Nov 6–13, 2026. The layer stack is a schematic, not a measurement.
Graphic: Powered by Thorsten Meyer AI
Why PDF Edge Cases Matter
Docestra’s development illustrates a practical gap between software that works in a demonstration and software that can be trusted with everyday documents. A save defect can erase annotations, while a cover-up mistaken for redaction can leave sensitive text in the file. The distinction is consequential for people editing contracts, forms or records.
Meyer’s account suggests that testing against varied real files, rather than relying only on a few examples, changed how the editor handled document restrictions and sensitive data. The reported test results are evidence about this project’s own checks, not independent proof that every PDF or redaction is safe. The article also does not establish that the tool saves time or money.
Popular right now
KitchenAid 3-Piece Pasta Roller and Cutter Set for Stand Mixers
TCL 65-Inch 4K UHD Smart LED TV (65S425)
TCL 55-Inch T7 Series 4K QLED Smart Google TV (55T7)
As an affiliate, we earn on qualifying purchases.
From One-Day Build to Version 1.5.1
Meyer describes “vibe coding” as building software mainly by giving an AI model plain-language instructions and having it write code. He says Claude Code, running Claude Opus, generated Docestra’s code, while OpenAI’s Codex reviewed it from the start. The application is a local TypeScript web app running in a browser on his Mac; he says files remain on his machine.
The first release already included viewing, search, annotations, page operations, printing and saving. Subsequent changes followed problems found in testing and features Meyer wanted for his own workflow. He says the PDF viewer library includes fonts under the GPL-2.0 licence, which he chose not to ship because the project’s rules exclude GPL code. He reports checking licences explicitly before adding later libraries for text recognition and editing.
“Vibe coding gets you a working tool quickly. Turning it into a tool you can rely on is the real work.”
— Thorsten Meyer, writing on ThorstenMeyerAI.com
Limits of the Reported Tests
The results are reported by the tool’s developer; the source does not provide an independent audit of the test methods, the 930- and 917-file collections, or the zero-leak redaction result. It also does not establish how Docestra performs across every PDF format or whether other users would encounter the same issues.
Meyer says he has not measured time or cost savings. Docestra is private and is not available to download, so readers cannot independently reproduce the reported tests using the application. Version 2.0 was still in development as of November 13, with no completion date given.
Version 2.0 Still in Development
Meyer says work on Docestra 2.0 began November 13, but the version is unfinished. His account does not give a release date or say whether the tool will become publicly available. For now, the next reported development is continued work on the private application, with its test suite and fixes forming part of the project’s ongoing process.
Key Questions
What is Docestra?
Docestra is Thorsten Meyer’s private PDF editor, built with AI coding agents. He says it runs locally in a browser on his Mac and is not available to download.
What problems did testing uncover?
Meyer reports a save issue that could make annotations disappear, restrictions affecting encrypted PDFs, and weaknesses in an early redaction design. He says these led to changes in saving, file unlocking and redaction.
How did the final redaction design work?
According to Meyer, the revised design converts pages with redactions into images, fills the marked areas in the image and adds a searchable text layer that omits lines touching a mark. He reports zero detected leaks in a 917-PDF test; that result has not been independently verified in the source material.
Has Meyer shown that Docestra saves time or money?
No. Meyer explicitly says he has not measured time or cost savings. He says testing and rework took longer than building for several packages.
When will Docestra 2.0 be released?
No release date is provided. Meyer says version 2.0 has been in development since November 13, 2026, and is not finished.
Source: ThorstenMeyerAI.com
