Riley Intelligence is a local-first Electron app for AI-assisted analysis and drafting. It runs on your machine with a Python backend and supports local models (Ollama) plus optional API-based providers.
- Electron desktop shell (React + Vite renderer)
- Local chat workflow with streaming responses
- File attachment upload + extraction
- Draft generation hooks for slides and prototypes
- Knowledge-base search interface (stubbed in this release)
- Data-provider integration interface (stubbed in this release)
This open-source cut removes organization-specific integrations. The following endpoints are placeholders and return actionable "not implemented" messages:
- POST /data_provider/test
- POST /data_provider/manifest/refresh
Implement your vendor logic in backend/data_provider_adapter.py and wire it into backend/main.py/chat.py as needed.
- Install dependencies
npm install- Run in development
npm run dev- Build desktop app artifacts
npm run build- Electron
- React + TypeScript + Vite
- FastAPI backend
- Ollama (optional local LLM runtime)
- electron/: main process and preload bridge
- renderer/: React frontend
- backend/: FastAPI services and model orchestration
- shared/knowledge_base/kb_search.py: minimal KB shim for v0.1
- Auto-update is disabled in this open-source v0.1 baseline.
- Keep credentials out of source control.