FastVC

FastVC is an AI workspace for venture investment teams. It adapts the proven PEHero workflow to a narrower startup-investing domain: thesis-led discovery, founder and fundraising signals, screening, round construction, diligence, investment committee, LP relations and portfolio support.

Product surface

• Startup discovery by thesis, sector, geography, stage and momentum
• Saved searches, alert cadence, company signals, founder profiles and warm paths
• Pipeline from discovered through Series C/growth investment and exit
• 25 venture-native agents across sourcing, ownership, diligence, IC/LP and portfolio
• Round, option-pool, dilution, pro-rata and probability-weighted outcome modelling
• Data-room upload and retrieval with cited source evidence
• LP/family-office CRM and LP-update workflows, without tax functionality
• BYOK adapter stubs for Affinity, Attio, Pipedrive and Brevo
• English plus the PEHero multilingual UI framework

The app uses the fastvc OLTP schema and fastvc_rag vector schema. Both are separate from PEHero while using the same PostgreSQL database configured by DB_URL.

Run locally

cp .env.example .env
python -m venv .venv
.venv/bin/python -m pip install -r requirements.txt
.venv/bin/python -m db.migrate
.venv/bin/python -m synthetic.generate --seed 42 --skip-rag
.venv/bin/python main.py

Open http://localhost:5059. The debug health check is http://localhost:5059/app/_debug/ping. The container and reverse-proxy health endpoint is http://localhost:5059/healthz.

Required configuration:

• DB_URL
• XAI_API_KEY for live agent responses
• APP_SECRET for sessions and per-user integration-key encryption

Optional configuration includes Tavily or Exa search, Google OAuth, Postmark, and global Affinity/Attio/Pipedrive/Brevo credentials. Users may instead add their own integration keys in /app/integrations; the current provider adapters validate and store configuration but do not initiate live sync.

Repository map

agents/       25 specialist agent definitions and intent routing
chat/         app workspaces: discovery, signals, pipeline, models, LPs
db/           idempotent fastvc and fastvc_rag schema migrations
prompts/      venture vocabulary plus specialist operating instructions
synthetic/    deterministic startup, founder, round and signal corpus
tools/        startup, round, diligence, LP and integration tools
landing/      public product pages
tests/        routing, agent construction, tools and UI smoke coverage

Production deployment

FastVC deploys from main to Coolify using the repository Dockerfile. The canonical production origin is https://vc.fastsme.com, the exposed container port is 5059, and the health path is /healthz. Set the variables listed in .env.coolify.sample; docker-entrypoint.sh runs the idempotent schema migration before starting the application. An authenticated GitHub Actions job automatically deploys updates to main after CI passes.

Safety and data boundaries

• Always schema-qualify SQL with fastvc.* or fastvc_rag.*.
• Treat generated screens and synthetic signals as research leads, not verified facts.
• Keep evidence and assumptions separate in investment outputs.
• Round and outcome models are illustrative and require legal, tax and fund-model review.
• Provider API keys are encrypted at rest and are never rendered back to the browser.