Welcome to rolv
rolv is your intelligent research companion — an agentic desktop app for data analysis, reproducible pipelines, and scientific computing.
Quick Links
- Getting Started — Login, your first chat, working directories, and sidebar basics
- Chat View — Conversational AI with code execution, file attachments, and auto-run
- Graph View — Build and execute DAG pipelines with compute, research, and data nodes
- Recipes — Reusable workflow templates: browse, import, edit, publish, and sync
- Reviews — Review community contributions to your recipes
- Settings — Token usage, execution tools, remote compute, custom LLM/embedding models
- Troubleshooting — Common issues and how to resolve them
Key Concepts
| Concept | Description |
|---|---|
| Session | A conversation (chat) or pipeline (graph) workspace |
| Graph | A directed acyclic graph (DAG) of processing nodes |
| Node | A single step in a graph — run code (compute), read/search and write a report (research), or manage files (data in/out). Research always writes {label}_report.md; optionally also a JSON table for the next step. |
| Recipe | A reusable, shareable pipeline template with code, environment, and documentation |
| Working Directory | The local folder where code runs and files are stored. Chosen when a disk tool needs it, not on every chat message |
Getting Help
- Use the Ask a Human Expert button (🎓) in chat or graph views to get scientific guidance
- Report software bugs on GitHub Issues