Fichero: For archives
Fichero is a Mac app to read and work with archives. It can help transcribe handwritten documents, or generate structured data from using Automatic Text Recognition (ATR). It can also help you navigate, search, and edit the results. To learn more about how Fichero works, see the user guide, which is coming soon.
Fichero is open source. It is in early days, 100% vibe coded, and in Public Alpha.
Features
-
Library
Documents are stored in a library, which can contain folders, images, pdfs, audio, video, and other documents. Browse them as lists, icons, tables, and columns; or extract data and browse them as datasets, timelines, and boards; or organize them visually on a canvas.
-
AI workflows
Use built-in AI tools and workflows (or chain both together) along with custom free local or cloud AI models, or make your own.
-
Preview
The image, PDF, and document viewer, lets you view a document; the word boxes over the scan; draw and save regions; make highlight and annotations; flip between different renditions of a page.
-
Edit
Rotate, crop, enhance, and remove backgrounds to make reading a document easier.
-
Reader
Keep transcripts, translations, and other “artifacts” that you or AI write or generate, side by side with the original image.
-
Inspector
Edit metadata, artifacts, entities, annotations, edit history, and the knowledge graph, for each document.
-
Knowledge
Automatically read names, places, and dates out of your documents and turns them into simple statements. (e.g. who did what, where, and when). Each statement is tied to the exact words on the page it came from. Fichero can then create a knowledge graph of your archive, which you can edit, or browse as biographies, maps, and timelines. Each “statement” is tied to its source text.
-
Search
Use keyword or semantic search powered by a powerful vector database to search across a library or a folder, and see the relevant text with matched words highlighted in the results.
Fichero is open source under AGPL-3.0.