runway-ai-usage-tracker
Local-first AI quota dashboard tracking remaining capacity and reset timers across multiple generative AI providers including Claude, Gemini, ChatGPT, GitHub Copilot, OpenRouter, MiniMax, Ollama, OpenCode, zAI, and more.
Ficha actualizada 22 sept 2026, 19:22 · upstream 22 sept 2026, 19:22 · release v2.12.0
Qué es runway-ai-usage-tracker
Descripción editorial construida a partir del repositorio, su documentación y sus señales de mantenimiento.
Runway is a local-first monitoring tool that tracks remaining capacity and reset timers for your entire generative AI stack. It aggregates usage data from multiple providers into a single dashboard with persistent history and fleet management capabilities. The application uses event-sourced architecture to maintain an immutable audit trail of all assistant messages, enabling precise cost tracking and forecasting. Runway supports 13+ providers through a multi-tier fallback collection strategy that prioritizes API calls over web scraping or local file parsing when available.
- Monitor remaining tokens across multiple AI models to prevent quota exhaustion
- Track token usage trends per provider to identify cost optimization opportunities
- Forecast resource depletion and plan budget allocations proactively
- Audit API consumption patterns for compliance and billing reconciliation
- Set up automated alerts when providers approach their rate limits
- Monitor session duration and composition across different AI assistants
- Supports 13+ AI providers with a unified dashboard interface
- Three-tier fallback collection strategy (API → Web scraping → Local files)
- Event-sourced immutable history for complete audit trail of all messages
- Smart caching with configurable intervals and automatic stretching after inactivity
- Theil-Sen regression forecasting on quota snapshots to predict exhaustion
- Persistent SQLite-backed usage snapshots with hourly resolution on short windows
- Provider detail pages include token trend charts, composition heatmaps, and session breakdowns
- Month-to-rolling time scope toggle for flexible reporting across any period
- Docker containers cannot access native desktop keychains for cookie-based authentication
- Cookie collectors require manual configuration per provider or via environment variables
- Web scraping fallback may fail if target websites block automated requests
- Local file parsing is limited to text content and does not support structured data extraction
- Rate limiting applies at 600 events per minute per IP address for sidecar ingestion endpoints
Perfil de infraestructura
Estimación inicial derivada del repositorio y de los servicios detectados.
Despliegue detectado
Qué señales reales de instalación hemos encontrado.
Runway requires Python 3.12+ and Node.js 18+ for the UI, with Docker providing a headless server runtime. Use docker-compose to run the full stack including persistent storage and environment variables. For cookie-based providers (Claude, ChatGPT), configure OAuth tokens or session cookies via environment variables. The sidecar agent script must be installed on each host to collect browser data. Use DB_ENCRYPTION_KEY for sensitive metadata protection in your SQLite database.
Stack y dependencias
Componentes inferidos a partir de manifests, configuración y documentación.
Puede sustituir a
Relaciones detectadas por el análisis del repositorio y su categoría.
Experiencia real de la comunidad
Ayuda a contrastar la estimación automática con instalaciones reales.
Todavía no hay datos comunitarios para este proyecto.