NeuroForge / Florra OS

The operating layer — an agent-run content and marketing machine for the Florra network. The official site is florra.net; this repo is the engine behind the "florra os" petal.

The loop: generate → publish → learn → scale. Agents run around the clock; a human makes the calls.

NeuroForge pipeline → Clip Engine → Postiz → OpenClaw/optimizer → dashboard → Notion desk

Quick start

# See the whole world
cat FLORRA_OS.md

# Generate content for a topic (needs ANTHROPIC_API_KEY)
cd neuroforge-os
python scripts/neuroforge_pipeline.py --topic "Stop Overthinking" --faculty "Dr. Nova Vale" --mode full

# Run the dashboard (works offline, no keys)
python florra_dashboard.py

# Manage the launch desk (local fallback until NOTION_TOKEN is set)
python notion_todo.py list
python notion_todo.py add "Host Postiz and connect accounts" --tag launch

Components

Component	Path	Needs
Pipeline (6 agents + QA)	neuroforge-os/scripts/neuroforge_pipeline.py	Anthropic key
Google ADK agents	neuroforge-agent/, florra-agent/	agents-cli, Gemini key
Clip Engine	clip-engine/ (merge via merge-clip-engine.sh)	Postiz key
Postiz posting	neuroforge-os/scripts/postiz_integration.py	Postiz host + accounts
Airtable sync	neuroforge-os/scripts/airtable_sync.py	Airtable key
Ad integrations	neuroforge-os/scripts/{meta_ads,tiktok_ads}_integration.py	ad account creds
Optimizer	neuroforge-os/scripts/optimize_prompts.py	Anthropic key
Weekly cron	.github/workflows/weekly-content-routine.yml	GitHub secrets
Dashboard	florra_dashboard.py	none (reads local files)
Notion desk	notion_todo.py	NOTION_TOKEN, NOTION_DB_ID

Env

Copy neuroforge-os/.env.example → .env and fill in keys. For GitHub Actions, set the same values as repo secrets.

Non-negotiables

• Honest AI labeling on every post.
• No fake citations, no medical claims, no overclaiming.
• Human review before publishing. Agents run around the clock; a human makes the calls.