Memory Store
Early-stage startup building a shared memory layer that lets teams and AI agents recall company knowledge.
The call
Memory Store sits in a newly named, crowded category where cloned startups and platform owners are moving fast, and no customer numbers are public yet.
Memory Store takes emails, chats, and documents from across your work apps and organizes them so AI assistants can remember context about your company and projects. The problem is that what Memory Store does technically is straightforward to copy, companies like Mem0 and Letta are already funded much better, and Claude and ChatGPT are building this memory feature directly into themselves rather than relying on outside tools. The one thing that could save Memory Store is if teams get so dependent on how it organizes their specific company's history that switching to something else becomes painful, but that lock-in has not been proven yet and probably will not happen before better-funded competitors or the big AI companies themselves capture this market.
Why
01
AI pressure: ChatGPT and Claude native persistent memory. ChatGPT and Claude already remember user and team context, cutting the reason to bolt on a separate memory layer. Could AI do the core job? Yes, likely.
02
Bigger players: Glean, Microsoft 365 Copilot, Google Workspace. Glean and Copilot already search and surface company knowledge across tools, doing a similar job at enterprise scale. Could a giant copy it? Yes, easily.
03
What protects it. Founders' Julep AI open-source project (7K+ stars) gives early developer trust and a distribution channel; little else stands out.
What would change this call
Anthropic or OpenAI ships native cross-app persistent memory tied to MCP. Would show Memory Store's core function absorbed into platform features it cannot control.
Memory Store discloses a named paying customer or design partner before its next Demo Day. Would show it can retain users despite lacking funding or scale of rivals.
Mem0, Letta, or MemPalace announce new funding, enterprise deals, or a merger/acquisition of a smaller peer. Would show consolidation pace and how quickly better-funded rivals pull ahead.
Memory Store discloses a named paying customer or design partner before its next Demo Day. Would show it can retain users despite lacking funding or scale of rivals.
Mem0, Letta, or MemPalace announce new funding, enterprise deals, or a merger/acquisition of a smaller peer. Would show consolidation pace and how quickly better-funded rivals pull ahead.
Key numbers
- Team of 4 people, San Francisco
- Public launch on YC Launch, March 2026
- Rival MemPalace hit 36,000+ GitHub stars in days
Named risks
- Native persistent memory in Claude or ChatGPT could commoditize Memory Store's core ingest-summarize-retrieve function.
- Crowded 'Company Brain' category with better-funded peers (Mem0 $24M, Letta $10M) and near-identical YC batch-mates.
- No disclosed revenue, customers, or benchmarks as of May 2026 launch, leaving traction unverified.
- Category-wide consolidation or shutdowns expected among undifferentiated entrants within six months per May 2026 analysis.
- Product documentation and pricing flagged as unclear by a third-party review in July 2026.
Rivals, as read in July 2026
| Rival | Direction | Evidence |
|---|---|---|
| Mem0 | losing to | Mem0 raised $24M Oct 2025 and published benchmarks (92.5 LoCoMo) Apr 2026; Memory Store discloses no funding or benchmarks |
| Hyperspell | contesting | Near-identical YC-adjacent 'Company Brain' positioning competing for same undifferentiated niche as of 2026 |
| Hyper | contesting | Hyper claims 50+ teams and up to $1k MRR (vendor claim, 2026); Memory Store discloses no comparable traction |
| Letta | losing to | Letta shipped memory-first coding agent 'Letta Code' with $10M backing in March 2026 |
| Claude/ChatGPT native memory | losing to | Native persistent memory features expanding through 2026, threatening standalone memory-layer thesis |
| Glean, Slack/Salesforce, Google Workspace, Microsoft 365/Copilot | losing to | Suite owners already hold underlying data and face no business-model barrier to shipping cross-tool memory |
Questions people ask
- Is Memory Store profitable?
- No revenue, ARR, or customer counts are disclosed as of the 2026-05-19 YC Launch post. The company uses a sales-led 'book a call' motion rather than self-serve signup, suggesting early pilot stage with no confirmed profitability data available.
- Could AI replace Memory Store?
- Yes, plausibly. Memory Store's core function, ingesting and summarizing communications into retrievable context, is the same capability Claude and ChatGPT are internalizing as native persistent memory features, per a 2026-05-12 analysis of the 'company brain' category.
- Who is Memory Store's biggest competitor?
- Mem0, which raised $24M in October 2025 from Basis Set, Peak XV, GitHub Fund and Kindred Ventures and has 41,000 to 59,000 GitHub stars, is the best-funded rival, alongside near-identical YC-adjacent peers Hyperspell and Hyper.
- Who owns Memory Store?
- Memory Store is an independent, privately held startup co-founded by Diwank Tomer and Ishita Jindal, backed by Y Combinator (batch P26) as of the 2026-05-19 YC Launch post, with no other disclosed investors, valuation, or funding amount.
- What is Memory Store's product?
- Memory Store ingests a company's Slack, Gmail, Granola and Fathom data plus Claude and Codex sessions, grouping conversations by project, person and decision into a persistent, permissioned memory exposed via the Model Context Protocol, per its site as of 2026-05-19.
- Does Memory Store have a moat?
- No exclusive data, regulatory barrier, or lock-in is evident as of 2026-05-19. Its proposed differentiation, shared human-and-agent memory across tools, remains an unverified roadmap claim rather than a demonstrated, defensible advantage against better-funded rivals.
Timeline
- 2026-05-19Memory Store publicly launches on YC Launch platform with no disclosed revenue or customers
- 2026-05-12Analysis piece names 'Company Brain' as a distinct funding thesis following YC Summer 2026 RFS
- 2026-04-02Letta ships 'Letta Code,' a memory-first coding agent
- 2026-04-01Mem0 publishes 'State of AI Agent Memory 2026' report with benchmark scores
- 2026-04Rival MemPalace gains over 36,000 GitHub stars within days
- 2026-03Memory Store's approximate public launch date, about two months before its YC Launch post
- 2026-02-24Report details Julep's Postgres/pgvector infrastructure underlying Memory Store's tech
- 2025-12Anthropic donates MCP to the Agentic AI Foundation (Linux Foundation), making it an open standard
- 2025-10Mem0 raises $24M Series A from Basis Set, Peak XV, GitHub Fund, Kindred Ventures
- 2025-09-08Julep AI, Memory Store's predecessor project, documented with 7,000+ GitHub stars
Revision history
27 July 2026
First assessed. Losing its protection. The text above is fixed as of this date and is not edited afterwards.
In the same band
Sources
- Jobs at Memory Store
- Julep (2025-09-08)
- Launch YC: Memory Store - Your company's brain (2026-05-19)
- What Is a "Company Brain" and Why Do I Keep Hearing About Them? (2026-05-12)
- Medium (Mem0 funding/stats)
- Best AI Agent Memory Frameworks in 2026 (2026-04-02)
- MemPalace, Mem0, and Persistent Context (2026-07-08)
- Best Mem0 Alternatives in 2026 (2026-04-08)
- Hyperspell: Your Company Brain
- Hyper: Company brain that powers your AI employees
- One shared memory for your teammates, and agents
- State of AI Agent Memory 2026 (2026-04-01)