MemoryOS
MemoryOS provides a connected memory layer that unifies AI session context across ChatGPT, Claude, Gemini, and major IDEs. Built on the open MCP standard, it serves individual developers and knowledge workers via a freemium subscription SaaS priced up to $19 per month.
- Company typePrivate
- Founded-
- Headquarters—
- Headcount—
- GTM typeB2C
- OfferingSoftware
What MemoryOS does
MemoryOS provides a connected memory layer that links AI sessions across ChatGPT, Claude, Gemini, Cursor, Claude Code, Windsurf, VS Code, and other MCP-compatible tools into a single, navigable graph. When a connected assistant concludes a session, it saves a compact summary called a dot; dots auto-link to related dots through shared topics, and on the Max tier an AI agent adds typed links with explicit reasoning. Recall combines full-text search, semantic similarity, and graph traversal, rendering results as a token-budgeted context block served back to the user's AI tool of choice. A flagship capability, Projects, converts a connected GitHub repository into a code graph spanning 15 programming languages so that AI tools receive architectural context without re-reading source files.
The product is delivered as self-serve, web-based SaaS with three pricing tiers: Basic (free, 50 dots, chat apps only), Premium ($9/month or $79/year, unlimited dots plus IDE support, deep dots, weekly digest), and Max ($19/month or $189/year, which adds the MemoryOS Agent, Projects, AI-generated typed links, large context blocks, and a Personal API). Authentication is OAuth 2.1 with PKCE; each account's dots live in an isolated database hosted on Render, with Cloudflare for DNS/network and Stripe for billing. There is no sales team, no marketplace presence, and no enterprise contracts; growth is product-led through the website and developer documentation.
The company is privately held, in early access, and operates as an independent entity. No revenue, headcount, funding rounds, investors, leadership team, founding date, or geographic incorporation data are disclosed in the source materials; a privacy policy last updated June 11, 2026 indicates active operation. Primary customer segments are software developers and AI power users who work across multiple assistants and IDEs, with knowledge workers and researchers as a secondary segment.
MemoryOS firmographics
Firmographics- Name
- MemoryOS
- Legal name
- MemoryOS
- Website
- https://thememoryos.com
- Company type
- Private
- Operating status
- Operating
- Short description
- MemoryOS provides a connected memory layer that unifies AI session context across ChatGPT, Claude, Gemini, and major IDEs. Built on the open MCP standard, it serves individual developers and knowledge workers via a freemium subscription SaaS priced up to $19 per month.
- Ownership category
- akta.pro rank
MemoryOS business model
Business model- GTM type
- B2C
- Offering type
- Software
- Cost components
- Technology or R&D, Infrastructure, Personnel, Operations, Marketing or Sales
Revenue model
- Subscription — Basic (Free): Free tier with 50 dots of memory, chat apps only (ChatGPT, Claude, Gemini), short served context (3–4 lines per dot). No credit card required. Intended to make the habit stick and drive upsell to Premium.
- Subscription — Premium ($9/mo or $79/yr): Unlimited dots, IDEs included (Cursor, Claude Code, VS Code, Windsurf), 4x served context with deeper graph recall, deep dots (full-session compactions), weekly memory digest. Billed monthly or annually via Stripe.
- Subscription — Max ($19/mo or $189/yr): All Premium features plus the MemoryOS Agent, Projects (codebase graphing), AI links between dots, huge context, and personal API token. Billed monthly or annually via Stripe. Includes all premium features plus advanced capabilities for heavy daily use.
Pricing tiers
| Model | Billing | Price |
|---|---|---|
| Freemium | Monthly | Basic: Free, 50 dots, chat apps only |
| Subscription | Monthly | Premium: $9/mo or $79/yr, unlimited dots, IDE support |
| Subscription | Monthly | Max: $19/mo or $189/yr, agent, projects, personal API |
Go-to-market motion1 record
Distribution channels1 record
Marketing channels3 records
MemoryOS product offering
Product offeringCore offering
MemoryOS provides a unified cross-tool memory layer for AI assistants and IDEs, built on the open Model Context Protocol (MCP) standard. It captures session summaries as compact 'dots' in a graph database, auto-links them through shared topics, and serves token-budgeted context blocks to connected AI tools (ChatGPT, Claude, Gemini, Cursor, Claude Code, VS Code, Windsurf) via OAuth 2.1 + PKCE-secured connectors. The product is sold as a subscription with progressively unlocked capabilities across free, Premium ($9/mo), and Max ($19/mo) tiers.
Product overview
MemoryOS is a unified cross-tool memory layer for AI assistants and IDEs, built on the Model Context Protocol (MCP) open standard. The product architecture consists of three subscription tiers—Basic (free), Premium ($9/month), and Max ($19/month)—that progressively unlock features for AI memory storage and retrieval. Basic provides 50 dots across chat apps; Premium extends to IDEs with unlimited dots, deep dot compaction, and 4x context; Max adds the MemoryOS Agent for graph organization, Projects for codebase graphing across 15 languages, AI-generated typed links between dots, and a Personal API. The core product stores short session summaries ('dots') in a graph database, connecting related memories through shared topics. Context is served as token-budgeted blocks to any connected tool via OAuth-secured MCP connectors. The system auto-links dots with typed reasoning and supports hybrid retrieval combining full-text search with semantic similarity.
Differentiator
Problem solved
Functional benefit
Products and services
- MemoryOS
Companies that use MemoryOS
Customer profileSegments2 records
Ideal customer profiles2 records
MemoryOS technology and API
TechnologyTechnology focussed Yes
API detail
- Has API
- Yes
- API docs
- API detail
Core technology
AI maturity
App detail
Integration8 records
AI capability6 records
Feature10 records
MemoryOS partnerships and signals
Strategic signalPartnerships
Ten partnerships are on record, tiered core, infrastructure and minor.
- OpenAI (ChatGPT)coreMemoryOS integrates with ChatGPT via the MCP (Model Context Protocol) custom connector. Users add the MemoryOS MCP URL in ChatGPT's Advanced settings (requires a paid ChatGPT plan), approve OAuth access once, and ChatGPT sessions save dots to and recall context from MemoryOS. The connector works on both desktop web and mobile ChatGPT apps after a single desktop setup.
- Anthropic (Claude, Claude Code)coreMemoryOS integrates with both Claude (claude.ai and desktop app) and Claude Code (CLI terminal sessions) via MCP. For Claude, users add the connector in Settings → Connectors. For Claude Code, one CLI command (claude mcp add) adds the server, followed by browser-based OAuth approval. All sessions — chat and terminal — can save dots automatically when the standing instruction is installed.
- Google (Gemini)coreMemoryOS integrates with Google Gemini (via Gemini CLI) through the MCP connector added in ~/.gemini/settings.json. Gemini web app does not yet accept custom connectors; the CLI path is the current integration route.
- Cursor (by Anysphere)coreMemoryOS integrates with Cursor IDE via MCP: users add the MCP server URL in Cursor Settings → MCP → Add new global MCP server, then authenticate via the sign-in prompt. When the standing instruction is added to .cursorrules, every coding session automatically stores dots on its own.
- Windsurf (by Codeium)coreMemoryOS integrates with Windsurf IDE via MCP: users add the server URL in Windsurf Settings → Cascade → MCP servers, then authenticate when prompted. Adding the standing instruction to .windsurfrules makes dot-saving automatic across sessions.
- Microsoft (VS Code + Copilot)coreMemoryOS integrates with VS Code (requires Copilot agent mode) via MCP: users add the HTTP MCP server via the MCP: Add Server command palette action. The setup can be committed to .vscode/mcp.json, enabling team-wide sharing with one commit. The standing instruction (in Rules) is served at the top of every recall in every connected tool.
- RenderinfrastructureRender provides hosting and storage for MemoryOS. Each account's dots are stored in their own isolated database on Render's infrastructure. Render keeps short-lived disk snapshots (~7 days) for disaster recovery.
- CloudflareinfrastructureCloudflare handles DNS and network routing for MemoryOS. All traffic flows through Cloudflare's network, and everything is encrypted in transit via HTTPS.
- StripecoreStripe handles all payment processing for paid plans (Premium and Max). Card details never touch MemoryOS servers — only Stripe's customer and subscription IDs are stored by MemoryOS.
- ResendminorResend handles transactional email delivery for MemoryOS: verification emails, password resets, and weekly memory digests.
Scale indicators3 records
Recent moves6 records
Expansion highlights5 records
MemoryOS competitors and assessment
Company assessmentDirect peers
- Mem: Mem is an AI-powered notes and self-organizing workspace that builds a personal knowledge graph from a user's inputs. It is the closest direct competitor to MemoryOS in the AI-memory-for-individuals category, with overlapping value proposition (persistent, searchable context) but a different delivery surface (notes vs. cross-tool MCP layer).
- Letta: Letta (formerly MemGPT) builds a long-term memory and state management layer for LLM agents. It directly overlaps MemoryOS's mission of giving AI persistent, structured memory, though Letta focuses on developer/agent builders while MemoryOS focuses on end-user cross-tool recall.
- Zep: Zep provides a long-term memory and context engineering service for AI assistants and agents. It is a direct peer in the AI-memory-as-infrastructure category, competing on retrieval, summarization, and persistence across sessions.
- Cognee: Cognee is an open-source memory layer for AI that builds a knowledge graph from user data and serves it as context to LLMs. It is highly comparable architecturally (graph-based memory, hybrid retrieval) and targets a similar developer/builder audience.
- Limitless: Limitless (formerly Rewind AI) builds a personal AI memory product that captures and recalls everything a user has seen or said. It overlaps MemoryOS on the persistent-personal-memory value proposition but relies on ambient capture rather than MCP-mediated tool integrations.
Emerging players
- Reflect: Reflect is a networked notes app with AI features and backlinking between notes, building a personal knowledge graph. It is an adjacent peer in the personal-knowledge-graph space with partial feature overlap on graph-based recall, but does not integrate across AI tools via MCP.
Broad incumbents
- Notion AI: Notion AI adds memory, Q&A, and context features across Notion's workspace product. It is a broad incumbent in the productivity/notes space with overlapping AI-memory functionality, but is locked inside the Notion surface and is not a cross-tool layer.
- ChatGPT Memory (OpenAI): OpenAI's native ChatGPT memory and Projects features give ChatGPT persistent context across sessions. It is a broad incumbent in the AI-memory category and the most direct competitive threat to MemoryOS, though it is vendor-locked to OpenAI's ecosystem.
- Microsoft Recall: Microsoft Recall captures on-screen activity on Copilot+ PCs and serves it as AI memory. It is a broad incumbent in personal-AI-memory, with deeper OS-level access but a narrower (Windows) and non-portable delivery surface compared to MemoryOS.
Others
- Pinecone: Pinecone is a managed vector database widely used as the retrieval backbone for long-term AI memory. It is an adjacent infrastructure provider that powers memory layers like Zep and could similarly underpin or compete with MemoryOS's retrieval stack.
Market position
Strengths5 records
Weaknesses5 records
Competitive moat4 records
Key risks7 records
Key highlights7 records
Customer concentration
MemoryOS social profiles
Digital presenceMemoryOS financial estimates
Financial estimateRevenue estimate
Valuation estimate
MemoryOS leadership team
Management profileNumber of profiles
MemoryOS funding detail
Funding detailFunding overview
Funding rounds
Investors
Funding detail is available on the Subscription and Enterprise plan.Contact sales →
MemoryOS M&A and investment
M&A and investmentM&A
Investments
M&A and investment is available on the Subscription and Enterprise plan.Contact sales →
Frequently asked questions about MemoryOS
What does MemoryOS do?
MemoryOS provides a unified cross-tool memory layer for AI assistants and IDEs, built on the open Model Context Protocol (MCP) standard. It captures session summaries as compact 'dots' in a graph database, auto-links them through shared topics, and serves token-budgeted context blocks to connected AI tools (ChatGPT, Claude, Gemini, Cursor, Claude Code, VS Code, Windsurf) via OAuth 2.1 + PKCE-secured connectors. The product is sold as a subscription with progressively unlocked capabilities across free, Premium ($9/mo), and Max ($19/mo) tiers.
Is MemoryOS a public or private company?
MemoryOS is a private company. It is classified as unknown and is currently operating.
When was MemoryOS founded?
MemoryOS was founded in -1.
How does MemoryOS make money?
Three revenue lines are on record. Subscription — Basic (Free) is the primary driver. The others are subscription — Premium ($9/mo or $79/yr) and subscription — Max ($19/mo or $189/yr).
Who are MemoryOS's main competitors?
Direct peers on record are Mem, Letta, Zep, Cognee and Limitless. Reflect is listed as an emerging player. Broad incumbents are Notion AI, ChatGPT Memory (OpenAI) and Microsoft Recall. Pinecone is listed as an others.
Does MemoryOS have an API?
Yes. MemoryOS offers a Personal API for Max plan users. The API provides endpoints for account management (GET /api/me), retrieving the full memory graph as JSON (GET /api/graph), semantic and full-text recall with context blocks (GET /api/recall?q=&limit=&hops=), and saving new dots (POST /api/save with title, body, source, tags). Authentication uses Bearer token (mospat_...) generated in the app under Profile. The token is shown once and only a fingerprint is stored. Revoking the token kills it instantly, and it stops working if the plan lapses. Dot deletion is deliberately not exposed via the token API to prevent accidental destruction from a leaked token. The API is available exclusively to Max plan subscribers. Developer documentation is at thememoryos.com/docs.