Your knowledge lives across all sorts of files. This release makes it easier to bring that material into Living Brain and turn it into context your agents can actually use.

What you can upload

The compiler supports a wide range of material, organised into four types:

PDFs and documents

PDF, DOC, DOCX, TXT, MD, JSON

Slide decks and presentations

PPTX

Spreadsheets and exported data

XLS, XLSX, CSV

Screenshots and images

BMP, GIF, HEIC, HEIF, JPEG, JPG, PNG, TIF, TIFF, WEBP

Up to 25 files per upload, 100 MB each. We're also expanding support for URL-based content, so a link works the same way a file does.

What happens after you upload

The compiler detects what you've given it and extracts the useful parts: text, tables, structure, and image content. We've also improved how it handles messy formatting and extraction failures, so a badly formatted document is far less likely to come through broken or incomplete.

The flow behind this is straightforward end to end: upload, detect the content, extract it, organise it inside Living Brain, then retrieve it the moment an agent actually needs it.

What this means day to day

Less manual preparation. More of your knowledge is available exactly when it matters, without you having to reformat it, retype it, or explain it again before an agent can use it.

Build your personalized AI Agents with a Living Brain via app.chatchat.com

More about uploading files to your Living Brain: https://support.chatchat.com/en/articles/16445402-add-information-to-your-living-brain

Data Compiler: Bring More of Your World Into Living Brain