A private AI that does real knowledge work.
Talk to it directly. Watch it fact-check a paper three different ways using only your local models. Every framework is visible and forkable.
The Platform
What It Does
Build
Visual framework editor. Wire AI capabilities together with a node-based canvas. Every block, connection, and variable is inspectable.
Verify
Fact-checking surface. Run claims through multiple sources, see the evidence chain, know what's supported and what isn't.
DARP
Provenance tracking. Every output carries a record of what influenced it — which models, which frameworks, which human decisions.
Mote
General-purpose AI chat. Direct conversation with your local models. No intermediary, no cloud dependency, no data leaving your machine.
Principles
Local-first
Your data stays on your machine. Models run locally. The cloud is an option, never a requirement.
Provenance-native
Every output is traceable. DARP tracks what influenced what, so you can always prove the chain of work.
Framework-centric
You don't invoke models. You invoke frameworks — visible, forkable workflows that wrap models with your logic and your constraints.
AGPL 3.0
Free to use, free to modify, free to host. The license ensures the platform stays open — if someone improves it, those improvements come back to everyone.