Cerebro Docs

Documentation site for the Gnosis Analytics platform, built with MkDocs Material.

Live site: docs.analytics.gnosis.io

---

Quick Start

pip install -r requirements.txt
mkdocs serve
# http://127.0.0.1:8000

Build & Deploy

mkdocs build --strict          # static site → site/
docker build -t cerebro-docs . # Docker image (Nginx)

Builds now also generate LLM-facing artifacts in site/:

• llms.txt -- curated public index
• llms-ctx.txt -- core docs and research corpus
• llms-ctx-full.txt -- full public corpus
• *.html.md mirrors for each page in MkDocs navigation

These files are generated from mkdocs.yml navigation and page metadata. Do not maintain them manually under docs/.

Auto-Update Scripts

Model catalogs, API endpoints, and dashboard metrics are auto-generated from live data sources. Sections between <!-- BEGIN/END AUTO-GENERATED --> markers are overwritten by these scripts.

python scripts/update_docs.py                  # regenerate from dbt manifest
python scripts/update_docs.py --dry-run        # preview changes

DUNE_API_KEY=xxx python scripts/sync_dune_queries.py   # sync 1,288 Dune queries
python scripts/sync_dune_queries.py --cache dune_cache.json  # use cached data

Structure

docs/
├── getting-started/     Platform overview, quickstart, architecture
├── data-pipeline/       Ingestion, crawlers, transformation
├── api/                 Authentication, endpoints, filtering, rate limits
├── models/              dbt model catalog (8 modules, 387 models)
├── esg-reporting/       ESG methodology, carbon footprint, data pipeline
├── mcp/                 MCP server tools, reports, setup
├── dashboard/           Sectors and configuration
├── developer/           Adding endpoints, models, scrapers
├── operations/          Infrastructure, deployment, monitoring
└── reference/           Env vars, glossary, data dictionary, Dune queries

Contributing

1. Edit markdown in docs/
2. Preview with mkdocs serve
3. Verify with mkdocs build --strict
4. Open a PR