Problem

CORE addresses the fragmentation and context loss created when people use multiple AI tools and agents. ChatGPT may remember a project while Cursor or Claude starts from scratch, forcing users to re-explain themselves, copy and paste background, monitor separate apps, synthesize what is happening, and manually feed agents the context they need. The economic pain is lost time and attention: instead of AI removing operational overhead, users become the coordinators and spend their day working for the tools.

The main use case is a continuous workflow across assistants and work applications. A developer can brainstorm in Claude Desktop, move into Claude Code or Cursor, and continue with the relevant project history without manually transferring context. The same problem applies to work spread across email, GitHub, Linear, Slack, meetings, and documents: CORE aims to make the agent start with the full picture rather than making the user reconstruct it every time.

Product / Service

CORE is positioned as a personal AI operating system built around a persistent, user-owned memory graph. It indexes information from email, meetings, GitHub, Linear, Slack, and assistant conversations into a temporal knowledge graph that preserves preferences, decisions, prior context, provenance, and change over time. That memory can be shared across ChatGPT, Claude, Cursor, and other AI applications, while the open-source and self-hostable model lets users keep the system and data under their control.

The product combines memory with execution. Its MCP-based Toolkit connects more than 50 applications and exposes more than 1,000 actions, while Tasks let a user describe work in plain English for CORE to plan, run, repeat, and report back on. Webhook triggers and gateways can make work proactive, and deployment is available through the hosted application or self-hosting on a local machine, server, or VPS. The benefit is not merely better answers; it is fewer context switches and the ability to delegate multi-step work across the tools where it actually happens.

Market

CORE competes in the emerging personal AI assistant, AI memory layer, AI-agent automation, and developer-productivity markets. Its closest conceptual competitors include Zep and its open-source Graphiti framework, which provide persistent context graphs for agents, and Mem0, which provides a persistent memory layer for AI applications. It also overlaps with personal-assistant products such as Vellum, Claude Cowork, ChatGPT, Lindy AI, Mem.ai, and Manus. CORE’s differentiation is the combination of a portable memory layer that works across assistants, an individual-user orientation, open-source and self-hosted deployment, temporal and explainable knowledge-graph storage, and direct action across many applications.

The company has meaningful early product traction but no disclosed revenue or customer count in the available evidence. Y Combinator lists CORE as active, in its Summer 2023 batch, with a three-person team; the public GitHub repository shows 1,919 stars and 183 forks, and the product has been presented with a hosted free tier alongside its open-source release. These signals support describing CORE as an early-stage, developer-led product with community adoption rather than confidently labeling it pre-revenue or claiming commercial scale.