The tool should be inspectable, extensible, and usable without locking essential workflows behind a proprietary gate.
Open source • local first • file based
Your knowledge should belong to you.
Knowforge is a free, open-source knowledge workspace built around common file formats, native folders, offline access, annotation, and local AI workflows.
Vendor agnostic by design: your notes, documents, folders, metadata, and annotations should remain usable outside the app.
Knowforge is not another knowledge silo.
It is designed as a native, file-based workspace where the application adds power without taking ownership away from the user.
Your core knowledge base should work on your own machine with no required cloud connection.
Common files and native folders are the foundation, so your knowledge can survive any single app.
The architecture should allow future modules without turning the core into a closed platform.
Files are the ground truth.
Knowforge is built around durable formats and native folders, with compatibility goals for the ecosystems knowledge workers already use.
Markdown-first notes
Note content is grounded in Markdown files, with room for richer editing experiences on top.
Obsidian compatibility
Designed around Obsidian-flavored Markdown syntax and compatibility with Canvas and Bases workflows.
Pandoc-friendly syntax
Support for Pandoc-style authoring helps notes become papers, docs, websites, and long-form outputs.
Nested YAML properties
Front matter can represent structured metadata, not just flat tags and titles.
Block transclusion
Search for any block, embed it elsewhere, and edit the embedded note without breaking context.
Flexible editing modes
Block mode, live mode, and split-column layouts support both structured thinking and fluent writing.
Annotation is not an add-on.
Knowledge work often begins in other documents. Knowforge treats annotation as a core system layer.
PDF annotation
Capture highlights, comments, source references, and extracted notes from PDFs.
- Research papers
- Manuals and reports
- Long-form references
Web page annotation
Bring web reading into the same knowledge flow instead of scattering notes across browser tools.
- Article highlights
- Source-aware clips
- Connected notes
DOCX annotation
Treat office documents as first-class sources for review, markup, and knowledge extraction.
- Draft feedback
- Meeting documents
- Policy and project files
AI without giving up locality.
Knowforge is designed for local LLM workflows, so AI can assist your thinking while your knowledge base remains local first.
Local model workflow
Integrate with local model runners such as Ollama or LM Studio for summarization, drafting, search assistance, and structured extraction.
Optional web access
Deploy web access when you want access anywhere, while preserving a local-first mental model and avoiding a cloud-only dependency.
Customizable without becoming complicated.
The workspace should feel personal, but not require a programming project just to organize your work.
Favorites
Keep important notes, folders, canvases, or views close at hand.
Smart folders
Create dynamic views from metadata, tags, folders, annotations, and saved searches.
Own ordering
Use folder, group, and view ordering that matches how you think, not just alphabetical sorting.
A modular foundation.
The core should stay dependable while future capabilities can grow as modules.
File layer
Native folders, Markdown, documents, metadata, and durable common formats.
Knowledge layer
Links, blocks, transclusion, properties, annotations, and smart organization.
Workflow layer
Editors, views, favorites, smart folders, review flows, and source-driven writing.
Expansion layer
Future modules for AI, publishing, collaboration, automation, and specialized workflows.
Knowforge principle: the app should make your knowledge more powerful without making it less yours.
Open by default. Local by default. Yours by default.
This landing page is ready to deploy as a static site. Replace this section with GitHub links, documentation links, screenshots, or an early-access call to action when ready.
- License: free and open source
- Storage: native folders and common file formats
- Mode: local first and offline capable
- AI: local LLM workflow with optional web access