About
Memory Store builds a shared company brain that turns Slack, Gmail, meeting, and AI-agent activity into connected, persistent memory. It sells to teams and their AI agents, differentiating through a unified memory layer that makes decisions and context accessible across both people and software agents.
Market
Memory Store competes in the persistent AI-agent memory and enterprise knowledge-context market. Unlike developer-oriented memory layers that primarily provide APIs or agent runtimes, it positions itself as a shared company brain: it collects workplace conversations and meeting insights, makes them available across teammates and MCP-compatible agents, and adds living Briefs that update as memories change. Its differentiation is therefore cross-tool, team-shared memory and workflow continuity rather than only single-user personalization or standalone vector/graph memory infrastructure.
Memory Store is best suited to AI-native startups and small knowledge-work teams, from individual users up to roughly 12-person teams, whose founders, engineering, product, sales, or operations leads need shared context across multiple AI tools and agents. Its strongest use cases are content, sales, team productivity, playbooks, and cross-session technical work.
At a Glance
Problem
Memory Store addresses the information bottleneck created by AI-enabled work. As teams and agents move faster, decisions, customer insights, project details, and technical context become scattered across Slack, Gmail, Granola transcripts, Claude and Codex sessions, and employees’ heads. Y Combinator frames the economics as a compounding coordination cost: AI can make a team “100x faster” while creating “100x more things to keep track of,” leaving memory—not raw execution—as the constraint on speed.
The killer use case is recovering the context behind work that has already happened: finding why a decision was made, what a customer asked for, what a teammate is doing, or how a technical problem was fixed, without hunting through threads or repeatedly explaining the same project to another person or agent. Memory Store targets the resulting lost time, duplicated work, and slower handoffs.
Product / Service
Memory Store is a shared, persistent memory layer for people and AI agents. It automatically pulls insights from tools such as Gmail, Slack, Granola, Claude, and Codex, synthesizes them into an up-to-date “company brain,” and organizes the information around projects, people, and decisions. Its live “Briefs” are intended to update as new memories are created, turning scattered activity into continuously refreshed decision logs, team updates, and customer-ask documents.
The service runs as an MCP connection inside the agents a team already uses, so a memory recorded in Claude can be recalled in Codex or another MCP-compatible client. Its core operations include automatic context loading at the start of a conversation, recording facts and decisions, semantic recall, and team sharing. The promised benefit is continuity across tools without requiring a new workspace or filing habit; the team plan is priced at $150 per user per month and includes shared memory, unlimited Gmail and Slack sync, role-based access, cross-team search, onboarding, and priority support.
Market
Memory Store competes in the emerging AI-agent memory and shared company-knowledge category: a layer between workplace systems and the growing set of AI assistants that need persistent context. A 2026 survey of agent-memory frameworks places Memory Store alongside Mem0, Supermemory, Membase, Zep, Graphiti, LangMem, Cognee, CrewAI, and LlamaIndex, while Memory Store’s own positioning emphasizes a “company brain” shared by teammates and agents. These are best understood as direct or adjacent competitors across managed agent memory, retrieval infrastructure, and AI-native knowledge management rather than as identical products.
The company appears to be at an early commercialization stage. Y Combinator lists it as an active Spring 2026 company founded in 2025, and its YC launch page describes a launch roughly two months before the August 1, 2026 reference date. It has a public paid team plan, a 30-day refund policy, a free-to-try individual offering, and a demo-led sales motion. The reviewed public sources disclose no customer count, revenue, ARR, or usage figures, so Memory Store cannot be characterized as demonstrably scaled or definitively pre-revenue; the strongest public traction signals are its YC status, launched product, and published pricing.
Founders & Leadership
Funding History
Y Combinator, Rebel Fund, Pioneer Fund
Recent News
Memory Store launched on Y Combinator’s Launch YC as a shared memory layer for companies, teammates, and AI agents. It collects insights from Slack, Gmail, Granola, and other tools, synthesizes them into a continuously updated company brain, and makes the context available to MCP-compatible agents through live “Briefs.”
In a Hacker News “Ask HN: What are you working on?” post, Memory Store described its beta MCP server for ingesting workplace context from Slack, email, and calendars and making it available to MCP-compatible agents. The post reported 35 user interviews, a 90-person survey, and early user feedback about using the product for investor communications and other contextual workflows.
Active Roles
0No active roles right now.
Get notified when they postBusiness Model
Memory Store appears to use a B2B, sales-led software model serving teams and their AI agents; its company profile directs prospects to book a call. Public evidence does not disclose specific pricing tiers or revenue terms.