You can deploy this model and start earning money today!

venkateshannabathina/
project_ningen

Name: project_ningen
Availability: InStock

Views

No views yet

Ningen (人間)

AI boardroom that turns a startup idea into requirements, product output, marketing assets, operations docs, and strategy docs.

What It Does

Ningen runs a 4-agent board meeting (CEO, CTO, CMO, COO), then executes downstream agents on the latest generated requirements.

Board meeting engine in board_meeting.py:
- Round 1: opening positions
- Client clarifications
- Round 2: reactions
- Conflict detection and resolution
- Final briefings + requirements generation
Execution agents:
- execute_cto.py: generates a build prompt and runs Claude Code to create product output
- execute_cmo.py: generates/publishes marketing artifacts
- execute_coo.py: generates execution_plan.md and risk_register.md
- execute_ceo.py: generates investor_pitch.md and executive_summary.md
Memory layer in memory_store.py: stores summaries of recent runs in outputs/memory_store.json

Current Entry Points

run_all.py: full autonomous pipeline (board meeting + CTO + CMO + COO + CEO)
app.py: Streamlit frontend (embedded UI) that talks to api_server.py
start.sh: convenience launcher for API + Streamlit together
dashboard/app.py: terminal-style board meeting flow
ningen_env.py: RL-style experiment runner and reward logging

Project Layout

text

project-ningen/
├── board_meeting.py
├── api_server.py
├── app.py
├── run_all.py
├── execute_cto.py
├── execute_cmo.py
├── execute_coo.py
├── execute_ceo.py
├── memory_store.py
├── ningen_env.py
├── start.sh
├── dashboard/
│   └── app.py
├── outputs/
└── requirements.txt

Setup

1) Install dependencies

bash

python3 -m venv venv
source venv/bin/activate
pip install -r requirements.txt

2) Create `.env`

At minimum:

OPENAI_API_KEY=your_openai_key

Optional (used by marketing publishing paths):

env

RESEND_API_KEY=your_resend_key
RESEND_FROM_EMAIL=you@domain.com
RESEND_SUBSCRIBER_EMAILS=user1@example.com,user2@example.com
DEVTO_API_KEY=your_devto_key
DISCORD_WEBHOOK_URL=https://discord.com/api/webhooks/...

3) Install Claude Code CLI (for CTO build path)

npm install -g @anthropic-ai/claude-code

Usage

Run everything in one command

python3 run_all.py "AI-powered invoicing assistant for freelancers"

Without an argument, it prompts for an idea interactively.

Run only the board meeting

python3 dashboard/app.py

Run web UI (Streamlit + Flask API)

./start.sh

Or manually:

streamlit run app.py

Run individual execution agents

bash

python3 execute_cto.py
python3 execute_cmo.py
python3 execute_coo.py
python3 execute_ceo.py

Run RL baseline experiment

python3 ningen_env.py

Outputs

Board meeting outputs are stored per generated product folder under outputs/:

outputs/<ProductName>/<ProductName>_requirements.json
outputs/<ProductName>/<ProductName>_briefings.json

Execution artifacts are written to outputs/ root, including:

index.html
cto_claude_prompt.txt
cold_email.md
devto_article.md
discord_post.md
reddit_post.md
execution_plan.md
risk_register.md
investor_pitch.md
executive_summary.md
memory_store.json
episode_log.json
rl_experiment.json

Notes

Several files in this repository are generated outputs and may change often.
Execution agents read the latest available requirements/briefings from outputs/.

venkateshannabathina/project_ningen