THIS REPO IS IN AN ALPHA STAGE* This repo is an idea based off of other open-source intelligence platforms, it's not intended to be original, it's intended to create a useful situational awareness of open data. Long Live OSINT

Meridian

Open-source global situational awareness and collaborative intelligence platform.

Monitor 150+ live data feeds — conflicts, aviation, maritime, weather, cyber threats, financial markets, and more — on an interactive map. Collaborate in real time with your team using Plan Mode.

Self-host in under 5 minutes. Bring your own LLM (OpenAI, Anthropic, Groq, or fully local via Ollama). All core data feeds are free and open.

---

Features

• Interactive Map — 64 layers across 8 categories powered by MapLibre GL JS (no Mapbox billing)
• 150+ Live Feeds — USGS, NOAA, NASA FIRMS, ACLED, OpenSky, AISHub, CISA KEV, GDELT, and more
• AI Analyst — Natural language queries against all live feeds (OpenAI / Claude / Groq / Ollama)
• Plan Mode — Real-time collaborative workspace: shared map, annotations, event timeline, task board, watch list
• Alerts — Plain-English alert rules with email, webhook, Slack, and Discord delivery
• Intelligence Reports — AI-generated daily briefs and on-demand situation reports

Quick Start

Prerequisites

• Docker and Docker Compose
• Git

1. Clone and configure

git clone https://github.com/your-org/meridian.git
cd meridian
cp .env.example .env

Edit .env and set your SECRET_KEY. Everything else is optional for a basic running instance.

2. Start

docker compose up

That's it. Visit http://localhost:5173.

• API docs: http://localhost:8000/docs
• API redoc: http://localhost:8000/redoc

3. Configure an LLM (optional)

AI features (chat, daily brief, situation reports) require a configured LLM provider. Edit .env and uncomment one of the provider blocks, then restart:

# Local (free, no API key):
LITELLM_PROVIDER=ollama
OLLAMA_MODEL=llama3.2

# Or OpenAI:
LITELLM_PROVIDER=openai
OPENAI_API_KEY=sk-...

---

Project Structure

meridian/
├── apps/
│   ├── api/        # FastAPI backend — REST API, feed workers, AI services
│   ├── web/        # React + TypeScript frontend — map, panels, UI
│   └── collab/     # Node.js Yjs WebSocket server — Plan Mode real-time sync
├── infra/
│   └── schema.sql  # PostgreSQL + PostGIS + TimescaleDB schema
├── docker-compose.yml
└── .env.example

Adding a Data Source

Every feed implements a single abstract base class. See apps/api/workers/base.py:

class FeedWorker(ABC):
    source_id: str
    category: FeedCategory
    refresh_interval: int  # seconds

    async def fetch(self) -> list[GeoEvent]: ...
    def health_check(self) -> FeedStatus: ...

See apps/api/workers/usgs_earthquakes.py for a complete example. Contributions of new data source workers are welcome — open a PR.

Tech Stack

Layer	Technology
Frontend	React 18 + TypeScript + Vite
Map	MapLibre GL JS
Styling	Tailwind CSS
Real-time	Socket.io
Collaboration	Yjs + y-websocket
API	FastAPI (Python 3.12)
Database	PostgreSQL 16 + PostGIS + TimescaleDB
Cache	Redis 7
AI	LiteLLM (OpenAI / Anthropic / Groq / Ollama)
Workers	APScheduler

License

AGPL-3.0 — free to use, modify, and self-host. Any modifications deployed as a service must be open-sourced.