Summary Pythia provides a unified, locally-run API delivering the planet's live state and forecasts, addressing information overload for AI agents and power users. It enables comprehensive global situational awareness without cloud costs or API keys, empowering informed decision-making.

How It Works Pythia fuses MiroFish (swarm-intelligence prediction) with Osiris (live global-intelligence globe), running entirely on local hardware via Ollama. Osiris ingests over 30 free, keyless live feeds (news, conflict, weather, markets, cyber). A local LLM drafts predictions (24h, week, month, year) with reasoning. A council of specialist agents re-evaluates these forecasts, surfacing consensus and dissent for nuanced, swarm-validated output.

Quick Start & Requirements

• Prerequisites: Ollama with a chat model (ollama pull llama3.1), Osiris checkout with PYTHIA overlay (integrations/osiris/INSTALL.md), Python 3.11+, uv.
• Run: Clone repo, copy .env.example to .env, run ./run-all.sh to start Osiris UI (:3000) and agent API (:8088).
• Docs: API reference at http://localhost:8088/docs.

Highlighted Details

• Swarm Deliberation: Forecasts are re-judged by specialist agents, surfacing consensus and dissent.
• Performance Tracking: Forecasts are graded against archived world state, updating a Brier scorecard; consensus is Brier-weighted.
• "What If" Scenarios: Supports counterfactual forecasting via /whatif API.
• Comprehensive Data Feeds: Ingests dozens of real-time, keyless feeds (conflict, disasters, markets, cyber, social).
• Agent API: Exposes a local HTTP+JSON API (/agent/view) for AI agents' global situational awareness.
• Local LLM Flexibility: Allows assigning different Ollama models to swarm personas or globally.

Maintenance & Community No specific details regarding maintainers, sponsorships, or community channels were found in the provided README content.

Licensing & Compatibility

• License: MIT.
• Compatibility: Designed for local execution and agent integration via HTTP API/MCP server. MIT license permits commercial use.

Limitations & Caveats

• Setup Complexity: Requires Ollama and a separate Osiris checkout with overlay.
• Resource Requirements: Running multiple LLMs and numerous live data streams locally demands significant computational resources.
• Data Dependency: Accuracy relies on the quality and availability of numerous third-party, keyless data feeds.