About Leo
We're building the search engine for human history
Leo
July 8, 2026
Most of the human record is handwritten but remains locked away in archives, largely unread and unsearchable. Transcription is the overlooked, labor-intensive bottleneck that gates access to billions of historical documents: deeds, accounts, inventories, minutes, diaries, letters, reports, and treatises. Leo turns them into searchable text, bringing document management, advanced search, and annotation together in one intuitive workflow. In doing so, it closes the gap between finding a source and understanding it.
Where traditional handwritten text recognition tools make users transcribe many pages manually to train a custom model before they get anything useful, Leo is built on the latest advances in machine learning and tuned for the complexities of historical handwriting, delivering high-accuracy results immediately.
History and physics
Leo was built to solve a problem every archival researcher knows firsthand. While researching his PhD dissertation, Jon Cooper spent long hours hand-transcribing Elizabethan and Jacobean manuscripts just so he could search them for a keyword, as the existing tools for automated transcription couldn't cope with the difficult secretary hand of these texts. Frustrated, he mentioned the problem to his friend Jack Weston. The two soon realized they were well placed to solve it together and founded Leo in late 2024.
Jon Cooper read History at Cambridge, earning a Starred First alongside an MSc in the History and Philosophy of Science, and is now a PhD candidate in History at Stanford. He studies the political economy of early modern Britain and its empire.
Jack Weston took a First in Physics at Oxford and a PhD in Physics at Cambridge, where his research developed new machine-learning methods for interpreting particle-physics data. He was previously co-founder and CTO of Novoic, an AI biotechnology startup.
Where we're headed
Our mission reaches beyond academic research. By making documents accessible, searchable, and interpretable, we aim to transform public engagement with the past—opening the world's archives to historians, archivists, genealogists, and dedicated amateurs alike, and drawing them, in time, into a single searchable record of human history.