Venture Analyst

Five AI agents dissect your startup idea in parallel - market, competition, risk, monetisation, synthesis - and return a scored verdict in real time.

---

Screenshots

---

How it works

Submit a one-line idea. Four specialist agents immediately fan out, each running concurrently and streaming their findings token by token. Once all four finish, a synthesis agent merges the outputs, scores the idea from 0–100, and issues a final verdict.

idea submitted
      │
      ├─── Market Research  ──┐
      ├─── Competitor Intel ──┤  (parallel)
      ├─── Risk Assessment  ──┤
      └─── Monetisation     ──┘
                               │
                         Synthesis
                               │
                         score + verdict

Events stream over SSE as they happen - thinking, tool calls, tokens, completion - so the UI updates live without polling.

---

Running locally

You need: Python 3.11+, uv, Node.js 18+, pnpm, and API keys for OpenAI and Tavily.

# 1. Backend
cd backend
cp .env.example .env          # add OPENAI_API_KEY and TAVILY_API_KEY
uv sync --group dev
uv run uvicorn main:app --reload

# 2. Frontend (new terminal)
cd frontend
pnpm install
pnpm dev

Open http://localhost:3000.

---

Agents

Agent	What it investigates	Searches the web
Market Research	TAM, trends, target users	✓
Competitor Intel	Top competitors, positioning gaps	✓
Risk Assessment	Legal, technical, execution risks
Monetisation	Revenue models, pricing channels
Synthesis	Merges all outputs → score 0–100

---

API

Three endpoints, nothing more:

POST /api/validate           { idea }  →  { job_id }
GET  /api/stream/{job_id}              →  SSE stream
GET  /api/result/{job_id}              →  scored result

Every SSE event is a flat JSON object:

{ "agent": "market_research", "type": "token", "data": "..." }

Event types in order: thinking → tool_call → tool_result → token → complete. The stream closes on system.done.

---

Stack

Backend - FastAPI · Uvicorn · Pydantic v2 · OpenAI · SSE-Starlette · Tavily · Loguru · Ruff · Mypy

Frontend - Next.js App Router · Turbopack · shadcn/ui · Tailwind CSS · Zustand · Framer Motion

---

Project layout

backend/
├── main.py                   ← FastAPI app + CORS + Loguru
├── api/routes.py             ← three endpoints
├── core/
│   ├── orchestrator.py       ← asyncio.gather fan-out
│   ├── streaming.py          ← SSE helpers
│   └── job_store.py          ← in-memory job registry
├── agents/
│   ├── base.py
│   ├── market_agent.py
│   ├── competitor_agent.py
│   ├── risk_agent.py
│   ├── monetisation_agent.py
│   └── synthesis_agent.py
└── schemas/                  ← Pydantic event / request / response models

frontend/
├── app/
│   ├── page.tsx              ← landing + idea input
│   ├── validate/page.tsx     ← live streaming dashboard
│   └── result/[jobId]/       ← final scored report
├── components/analysis/      ← AgentOutputCard, ActivityFeed, SynthesisOutput …
├── components/report/        ← ValidationScore
├── hooks/useAgentStream.ts   ← EventSource → Zustand
└── store/agentStore.ts       ← global state

---

Linting

# backend - ruff + mypy in one shot
cd backend && uv run python lint.py

# frontend - ESLint
cd frontend && pnpm lint