Initial commit — GCP Dot self-hosted clone

Server (FastAPI + SQLite) runs Stouffer Z network variance analysis.
Egg container uses os.urandom for hardware-entropy 200-bit trials.
Gitea Actions workflow auto-builds and publishes both Docker images.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

README.md

@@ -0,0 +1,109 @@
+# Global Consciousness Project — Self-Hosted
+
+A live network-variance tracker inspired by the [Global Consciousness Project](https://global-mind.org).
+Distributed "egg" containers contribute random numbers; the server runs statistical analysis and displays a colored dot.
+
+## Architecture
+
+```
+Egg containers (anyone can run)
+        │  POST /api/data  (one 200-bit trial per second)
+        ▼
+    Server (FastAPI + SQLite)
+        │  runs Stouffer Z network variance analysis every 60 s
+        ▼
+    Website  →  animated dot  +  history chart  +  embeddable iframe
+```
+
+## Quick start — run the server
+
+```bash
+git clone <this-repo>
+cd gcp
+docker compose up -d
+```
+
+The site is now at **http://localhost:8000**.
+
+To expose it publicly, put it behind a reverse proxy (nginx, Caddy) or deploy to any VPS / cloud service.
+
+## Contribute an egg
+
+Anyone can run an egg — one Docker command:
+
+```bash
+docker run -d --restart unless-stopped \
+  -e SERVER_URL=https://your-domain.com \
+  -v gcp_egg_data:/data \
+  git.hexadual.io/rocobo/gcp-dot-egg:latest
+```
+
+The egg:
+- Reads 200-bit trials from the OS hardware-entropy pool (`os.urandom`) once per second
+- Sends the trial count (0–200) to the server
+- Persists its ID across restarts via `/data/egg_id`
+
+## Images (built automatically by Gitea Actions)
+
+Every push to `main` builds and publishes both images to the Gitea container registry:
+
+| Image | Pull command |
+|-------|-------------|
+| Server | `docker pull git.hexadual.io/rocobo/gcp-dot-server:latest` |
+| Egg    | `docker pull git.hexadual.io/rocobo/gcp-dot-egg:latest` |
+
+## How the analysis works
+
+Every 60 seconds the server analyses the past hour of data:
+
+1. **Normalise** each trial to a Z-score: `z = (trial − 100) / √50`
+   (expected mean = 100, variance = 50 for Binomial(200, 0.5))
+
+2. **Stouffer Z** per second across all active eggs: `S_t = Σzᵢₜ / √N`
+
+3. **Network variance**: `V = Σ S_t²`  — follows χ²(T) under H₀
+
+4. **Index** = lower-tail CDF × 100:
+   - **High index (> 95)** → small variance → eggs are more coherent than chance → **blue**
+   - **Low index (< 5)** → large variance → eggs are noisier than chance → **red**
+   - **Middle (40–90)** → normal random behavior → **green**
+
+### Color table
+
+| Color  | Index     | Meaning |
+|--------|-----------|---------|
+| Blue   | > 95%     | Significantly small variance — deep coherence |
+| Cyan   | 90–95%    | Small variance — probable coherence |
+| Green  | 40–90%    | Normal random behavior |
+| Yellow | 10–40%    | Slightly elevated variance |
+| Orange | 5–10%     | Strongly elevated variance |
+| Red    | < 5%      | Significantly large variance |
+
+## Embed the dot
+
+```html
+<iframe src="https://your-domain.com/gcp.html"
+        height="48" width="48" scrolling="no" frameborder="0"></iframe>
+```
+
+## API
+
+| Endpoint | Description |
+|----------|-------------|
+| `POST /api/data` | Submit a trial `{egg_id, timestamp, trial}` |
+| `GET /api/status` | Latest analysis result |
+| `GET /api/history?limit=60` | Last N analysis records |
+| `GET /api/eggs` | Eggs active in the last 2 minutes |
+
+## Environment variables
+
+### Server
+| Variable | Default | Description |
+|----------|---------|-------------|
+| `DB_PATH` | `/data/gcp.db` | SQLite database path |
+
+### Egg
+| Variable | Default | Description |
+|----------|---------|-------------|
+| `SERVER_URL` | `http://localhost:8000` | Server to send data to |
+| `EGG_ID` | auto-generated | Override the egg's identifier |