Internal intelligence platform for DocMap. It is a monorepo of tools that started as manual ops - talking to patients and clinics, reading reviews, watching TikTok comments, drafting outreach by hand - and were productised so the same judgement scales.
Live web preview: intelligence-os-web.vercel.app
Hosted MCP: https://mcp.docmap.co.uk/mcp
Two layers:
- Platform packages - Next.js clinic UI, MCP, marketing ETL, workers (run in production)
- Absorbed project folders - earlier standalone tools kept here for convenience; several also have their own GitHub repos for clear ownership and READMEs
| Path | Role |
|---|---|
| App / UI (Next.js root) | Clinic outreach / accounts dashboard |
mcp-server/ |
Hosted MCP for Claude (TikTok tools, search, Gmail drafts) |
marketing-pipeline/ |
Production TikTok ETL + Supabase sync |
data-worker/ |
Railway cron jobs |
ingestion-pipeline/ |
Clinic CSV -> Supabase + embeddings |
gtm-pipeline/ |
GTM / CQC account intelligence |
relationship-desk-mcp/ |
Gmail relationship memory MCP |
docs/ |
Deploy, MCP onboarding, execution plans |
These folders are the "we did this manually, then automated it" toolkit:
| Folder in this repo | Standalone repo | What it is |
|---|---|---|
Clinic sales agent/ |
clinic-sales-agent | Doctify discovery, enrichment, CQC match, outreach drafts |
Carousel agents V2/ |
carousel-agent | Selection-first social carousels from documents |
Doctors Sales Agent/ |
doctor-sales-agent | Doctor onboarding email drafts from WhatsApp recommendations |
Negative Review analysis/ |
negative-review-analysis | Google negative-review themes for clinics |
Social media analysis/ |
social-media-analysis | TikTok metrics / transcripts / performance hypotheses |
Related product (separate org repo): synaptic-docmap/triage_tool - WhatsApp / specialist triage for patients. Ranking research: ranking_algo.
Living status: STATUS.md. Master plan: docs/MASTER_PLAN.md.
npm install
cp .env.example .env.local
npm run devcd marketing-pipeline
pip install -e .
python -m marketing_pipeline tiktok extract-components
python -m marketing_pipeline tiktok sync-supabaseSee docs/MCP_ONBOARDING.md and docs/DEPLOY.md.
Copy .env.example to .env.local:
MODEL_OCR- vision OCR (default Gemini Flash)MODEL_COMPONENTS- component cards (default DeepSeek flash)
| Doc | Topic |
|---|---|
STATUS.md |
Prod health and next steps |
docs/DEPLOY.md |
Railway / Vercel / MCP |
docs/MCP_ONBOARDING.md |
Connect Claude to DocMap MCP |
docs/MASTER_PLAN.md |
Phased product plan |