Moss

Moss provides sub-10ms semantic search for conversational AI, voice agents, and copilots, solving the problem of slow retrieval that breaks real-time AI systems. It targets developers and engineering teams building production AI systems, from startups to enterprises, who need low-latency retrieval. Unlike vector databases, Moss runs locally in the browser, edge, device, or cloud, eliminating network hops and latency bottlenecks, and integrates with existing LLM stacks like LangChain and Vercel AI SDK.

Key features

• Sub-10ms end-to-end retrieval latency
• Up to 100x faster than vector databases
• 100% local execution
• Offline indexing and querying
• No external vector database required
• Works with LangChain and Vercel AI SDK
• Runs in browser, edge, device, or cloud
• Real-time context retrieval for voice AI