| title | Tutorials |
|---|---|
| description | A guided learning path that takes you from understanding VisionClaw to running it, building your first 3D knowledge graph, and promoting a note into the formal ontology. |
| category | tutorial |
| difficulty-level | beginner |
VisionClaw Docs · Tutorials
These four tutorials form a single learning path. Work through them in order: each one assumes what the previous one taught. You start by understanding what VisionClaw is, install and run the stack, build a graph you can walk around, then enrich that graph by promoting a plain note into the formal ontology. No prior knowledge of RDF, CUDA, or actor systems is required — every step is explained as you reach it. There is no separate database to set up; the graph store is an embedded Oxigraph triple store running in-process inside the Rust backend.
By the end you can point VisionClaw at your own notes, navigate the result in 3D, and govern how concepts become first-class ontology classes.
| # | Tutorial | What you learn | Time |
|---|---|---|---|
| 1 | What is VisionClaw? | The mental model — what each part does and where your data flows, no commands typed | ~10 min |
| 2 | Installation | System requirements, Docker Compose setup, and optional NVIDIA GPU acceleration | ~20 min |
| 3 | Build Your First Graph | Start the stack, load notes from GitHub or JSON, navigate the 3D space, and enable GPU physics | 15–45 min |
| 4 | Promote a Note to the Ontology | Turn a Markdown page into a governed OWL class with SHACL validation and PROV-O provenance | ~30 min |
- Docker Engine with the Compose plugin (or Docker Desktop)
- 8 GB RAM minimum, 16 GB recommended
- A modern WebGL browser (Chrome, Firefox, or Safari, current release)
- Optional: an NVIDIA GPU with CUDA for the 55x physics speedup — VisionClaw runs without it on CPU
Tutorial 2 covers each of these in detail, including how to verify your setup.
Once you have finished the path:
- Pick a task from the How-To Guides — deployment, agent orchestration, REST API usage, XR setup.
- Understand the design in Explanation — architecture, the physics engine, the ontology pipeline.
- Look things up in the Reference — REST and WebSocket APIs, the binary protocol, configuration, and the agents catalogue.
- How-To Guides — task recipes once VisionClaw is running
- System Overview — how the pieces fit together
- Reference Hub — exhaustive API and configuration details
- Troubleshooting — fixes for common setup problems
- Governing ADRs: ADR-090 Hexagonal crate modularisation, ADR-112 Ontology augmentation retrieval spine