OneRingAI
Provides a unified API for multiple AI vendors' text, image, video, embedding, speech, and realtime voice models, with lifecycle-aware registries, realtime voice support, and a memory subsystem.
A stable connector-first API that unifies access to current OpenAI, Anthropic, Google, and xAI models across text, image, video, embedding, speech, and realtime voice. It solves the problem of integrating multiple AI vendors by providing a single TypeScript library with lifecycle-aware model registries, realtime voice support, and a memory subsystem. It targets developers and engineering teams building AI agents and applications, particularly in mid-market and enterprise companies. It is delivered as an npm package with a production-ready foundation, including connectors, permissions, memory, tools, orchestration, resilience, storage, and MCP.
Key features
- Connector-first agents
- Registry v2 model intelligence
- Realtime voice and media
- Lifecycle-aware model aliases
- Graph-native knowledge memory
- Vector and ranked retrieval
- Signal ingestion for profiles
- Standalone or agent-ready memory
- MCP support
- Typed events and tools
- No social media activity within the last 30 days
- Blog
- Newsletter
- Marketplace
- API
- Docs
- Changelog
- Software developers
- Engineering teams
- Enterprises