Lead Machine Lite

Landing page publica com diagnostico empresarial automatico via IA. O lead preenche um formulario, o sistema conduz um chat de ate 10 perguntas geradas pelo Claude e entrega um relatorio completo com estado atual, gargalos e futuro com IA.

Voce ve tudo no dashboard local e chega na call ja sabendo o perfil do lead.

Como funciona

Lead acessa URL publica
       |
       v
Preenche nome / email / empresa
       |
       v
Claude faz ate 10 perguntas personalizadas (FastAPI + Anthropic SDK)
       |
       v
Relatorio gerado: estado atual, gargalos, futuro com IA
       |
       v
Voce ve no dashboard: http://127.0.0.1:8792/dashboard

Stack

• Backend: Python 3.10+ + FastAPI + Uvicorn
• LLM: Claude API (Anthropic SDK) — usa sua propria chave, sem mensalidade
• Tunnel: Cloudflare Tunnel (trycloudflare.com) — URL publica gratuita
• Storage: JSON local em ~/zx-leads/ — dados 100% seus
• Rate limiting: slowapi

Instalacao (Windows)

Pre-requisitos: Python 3.10+, cloudflared instalado

# 1. Clone e instale dependencias
git clone https://github.com/Dimitrearaujo/lead-machine-lite
cd lead-machine-lite
python -m venv venv
venv\Scripts\pip install -r requirements.txt

# 2. Configure o .env
copy .env.example .env
# Edite o .env com sua ANTHROPIC_API_KEY e gere um ALUNO_TOKEN

# 3. Inicie
powershell -ExecutionPolicy Bypass -File scripts\start.ps1

Configuracao (.env)

ANTHROPIC_API_KEY=sk-ant-...        # console.anthropic.com/settings/keys
PORT=8792
LEADS_DIR=~/zx-leads
MODEL=claude-sonnet-4-6
ALUNO_TOKEN=                        # gere com: python -c "import secrets; print(secrets.token_urlsafe(24))"

Endpoints

Metodo	Rota	Descricao
GET	/health	Status do backend
POST	/lead/new	Cria lead e retorna primeira pergunta
POST	/lead/answer	Envia resposta e recebe proxima pergunta
GET	/lead/status/{id}	Status do diagnostico
GET	/lead/result/{id}	Relatorio final
GET	/dashboard	Dashboard do operador (requer ALUNO_TOKEN)
GET	/api/leads	Lista leads (requer ALUNO_TOKEN)
POST	/lead/generate-copy	Gera copy de marketing para o lead
POST	/lead/generate-kit	Gera kit comercial completo para o lead

Scripts Windows (pasta scripts/)

scripts\start.ps1    # sobe backend + tunnel, salva URL em tunnel-url.txt
scripts\stop.ps1     # para tudo
scripts\restart.ps1  # reinicia
scripts\status.ps1   # mostra estado, URL e ultimos logs

Nota sobre Windows

O fcntl.py na raiz e um shim para compatibilidade — substitui o modulo Unix fcntl por no-ops seguros para uso em servidor local single-process.

Licenca

MIT

---

🇺🇸 English

Lead Machine Lite

Public landing page with automated AI business diagnosis. The lead fills a form, the system conducts a chat of up to 10 questions generated by Claude and delivers a complete report with current state, bottlenecks and future with AI.

You see everything in the local dashboard and join the sales call already knowing the lead's profile.

How it works

Lead accesses public URL
       |
       v
Fills name / email / company
       |
       v
Claude asks up to 10 personalized questions (FastAPI + Anthropic SDK)
       |
       v
Report generated: current state, bottlenecks, future with AI
       |
       v
You see it on dashboard: http://127.0.0.1:8792/dashboard

Stack

• Backend: Python 3.10+ + FastAPI + Uvicorn
• LLM: Claude API (Anthropic SDK) — uses your own key, no subscription
• Tunnel: Cloudflare Tunnel (trycloudflare.com) — free public URL
• Storage: Local JSON at ~/zx-leads/ — 100% your data
• Rate limiting: slowapi

Installation (Windows)

Prerequisites: Python 3.10+, cloudflared installed

# 1. Clone and install dependencies
git clone https://github.com/Dimitrearaujo/lead-machine-lite
cd lead-machine-lite
python -m venv venv
venv\Scripts\pip install -r requirements.txt

# 2. Configure .env
copy .env.example .env
# Edit .env with your ANTHROPIC_API_KEY and generate an ALUNO_TOKEN

# 3. Start
powershell -ExecutionPolicy Bypass -File scripts\start.ps1

Key endpoints

Method	Route	Description
GET	/health	Backend status
POST	/lead/new	Creates lead and returns first question
POST	/lead/answer	Sends answer and receives next question
GET	/lead/status/{id}	Diagnosis status
GET	/lead/result/{id}	Final diagnosis report
GET	/dashboard	Operator dashboard (requires ALUNO_TOKEN)
GET	/api/leads	Lists leads (requires ALUNO_TOKEN)
POST	/lead/generate-copy	Generates marketing copy for the lead
POST	/lead/generate-kit	Generates full commercial kit for the lead

Windows scripts (scripts/ folder)

scripts\start.ps1    # starts backend + tunnel, saves URL to tunnel-url.txt
scripts\stop.ps1     # stops everything
scripts\restart.ps1  # restarts
scripts\status.ps1   # shows state, URL and latest logs

Note on Windows

The fcntl.py at the root is a compatibility shim — replaces the Unix fcntl module with safe no-ops for single-process local server usage.

License

MIT

---

← Back to profile