RuView Quickstart
Get a newcomer from zero to a running RuView sensing dashboard. Three tiers, pick the one that matches the hardware on hand.
Tier 0 — Docker, no hardware (2 minutes)
docker pull ruvnet/wifi-densepose:latest
docker run -p 3000:3000 ruvnet/wifi-densepose:latest
# open http://localhost:3000 — simulated CSI, full UI
Use this to demo the dashboard, explore the API, or develop UI without a sensor.
Tier 1 — Build the repo from source
# Rust workspace (1,400+ tests, ~2 min)
cd v2
cargo test --workspace --no-default-features
# Single-crate sanity check (no GPU)
cargo check -p wifi-densepose-train --no-default-features
# Python proof (deterministic SHA-256 pipeline check)
cd ..
python archive/v1/data/proof/verify.py # must print VERDICT: PASS
If verify.py fails on a hash mismatch after a numpy/scipy bump:
python archive/v1/data/proof/verify.py --generate-hash
python archive/v1/data/proof/verify.py
Tier 2 — Live sensing with an ESP32-S3 ($9)
This is the real thing. Hand off to the ruview-hardware-setup skill for the flash/provision/monitor loop, then:
# Lightweight sensing server (consumes the ESP32 UDP CSI stream)
cd v2
cargo run -p wifi-densepose-sensing-server
# Live RF room scan / SNN learning helpers:
node ../scripts/rf-scan.js --port 5006
node ../scripts/snn-csi-processor.js --port 5006
What to know before you start
- ESP32-C3 and the original ESP32 are NOT supported — single-core, can't run the CSI DSP pipeline. Use ESP32-S3 (8MB or 4MB) or ESP32-C6.
- A single ESP32 has limited spatial resolution — 2+ nodes (or add a Cognitum Seed) for good results.
- Camera-free pose accuracy is limited (~84s to train, modest PCK). For 92.9% PCK@20 use camera-supervised training (see
ruview-model-training skill, ADR-079).
- No cloud, no internet, no cameras required — everything runs on edge hardware.
Next steps to suggest
| Goal |
Skill / command |
| Flash & provision an ESP32 node |
ruview-hardware-setup · /ruview-flash · /ruview-provision |
| Tune channels / MAC filter / edge modules |
ruview-configure |
| Run a sensing application (presence, vitals, pose, sleep, MAT) |
ruview-applications · /ruview-app |
| Train a pose / sensing model |
ruview-model-training · /ruview-train |
| Multistatic mesh, tomography, cross-viewpoint fusion |
ruview-advanced-sensing · /ruview-advanced |
| Verify the build + generate a witness bundle |
ruview-verify · /ruview-verify |
Reference
README.md — feature matrix, hardware table, install options
docs/user-guide.md, docs/wifi-mat-user-guide.md, docs/build-guide.md, docs/TROUBLESHOOTING.md
docs/tutorials/, examples/ — runnable examples (environment, medical, sleep, stress, ruview_live.py)
1---
2name: ruview-quickstart
3description: Onboarding and first-run for RuView (WiFi-DensePose) — Docker demo with simulated data, repo build, and the fastest path to a live sensing dashboard. Use when someone is new to RuView or wants the shortest path to "it works on my machine".
4allowed-tools: Bash Read Write Edit Glob Grep
5metadata:
6 source: https://skillmd.com/skills/gabrielmoreira/ruview-quickstart
7---
8
9# RuView Quickstart
10
11Get a newcomer from zero to a running RuView sensing dashboard. Three tiers, pick the one that matches the hardware on hand.
12
13## Tier 0 — Docker, no hardware (2 minutes)
14
15```bash
16docker pull ruvnet/wifi-densepose:latest
17docker run -p 3000:3000 ruvnet/wifi-densepose:latest
18# open http://localhost:3000 — simulated CSI, full UI
19```
20
21Use this to demo the dashboard, explore the API, or develop UI without a sensor.
22
23## Tier 1 — Build the repo from source
24
25```bash
26# Rust workspace (1,400+ tests, ~2 min)
27cd v2
28cargo test --workspace --no-default-features
29
30# Single-crate sanity check (no GPU)
31cargo check -p wifi-densepose-train --no-default-features
32
33# Python proof (deterministic SHA-256 pipeline check)
34cd ..
35python archive/v1/data/proof/verify.py # must print VERDICT: PASS
36```
37
38If `verify.py` fails on a hash mismatch after a numpy/scipy bump:
39```bash
40python archive/v1/data/proof/verify.py --generate-hash
41python archive/v1/data/proof/verify.py
42```
43
44## Tier 2 — Live sensing with an ESP32-S3 ($9)
45
46This is the real thing. Hand off to the `ruview-hardware-setup` skill for the flash/provision/monitor loop, then:
47
48```bash
49# Lightweight sensing server (consumes the ESP32 UDP CSI stream)
50cd v2
51cargo run -p wifi-densepose-sensing-server
52# Live RF room scan / SNN learning helpers:
53node ../scripts/rf-scan.js --port 5006
54node ../scripts/snn-csi-processor.js --port 5006
55```
56
57## What to know before you start
58
59- **ESP32-C3 and the original ESP32 are NOT supported** — single-core, can't run the CSI DSP pipeline. Use ESP32-S3 (8MB or 4MB) or ESP32-C6.
60- A **single ESP32** has limited spatial resolution — 2+ nodes (or add a Cognitum Seed) for good results.
61- Camera-free pose accuracy is limited (~84s to train, modest PCK). For 92.9% PCK@20 use camera-supervised training (see `ruview-model-training` skill, ADR-079).
62- No cloud, no internet, no cameras required — everything runs on edge hardware.
63
64## Next steps to suggest
65
66| Goal | Skill / command |
67|------|-----------------|
68| Flash & provision an ESP32 node | `ruview-hardware-setup` · `/ruview-flash` · `/ruview-provision` |
69| Tune channels / MAC filter / edge modules | `ruview-configure` |
70| Run a sensing application (presence, vitals, pose, sleep, MAT) | `ruview-applications` · `/ruview-app` |
71| Train a pose / sensing model | `ruview-model-training` · `/ruview-train` |
72| Multistatic mesh, tomography, cross-viewpoint fusion | `ruview-advanced-sensing` · `/ruview-advanced` |
73| Verify the build + generate a witness bundle | `ruview-verify` · `/ruview-verify` |
74
75## Reference
76
77- `README.md` — feature matrix, hardware table, install options
78- `docs/user-guide.md`, `docs/wifi-mat-user-guide.md`, `docs/build-guide.md`, `docs/TROUBLESHOOTING.md`
79- `docs/tutorials/`, `examples/` — runnable examples (environment, medical, sleep, stress, `ruview_live.py`)