Memwyre
Memwyre is the AI Memory Platform that gives every AI tool, agent, and conversation a persistent, shared memory.
Is this your product? Claim it to edit the listing, post updates and answer questions. Prove it with an email address at the product's domain (the listing becomes Claimed), or put the LotsDirectory badge on your website (the listing becomes Verified, and its link to your website is followed while the badge is live).
Description
Memwyre is the AI Memory Platform that gives every AI tool, agent, and conversation a persistent, shared memory. Connect to Cursor, VS Code, ChatGPT, Claude, and more.
About Memwyre
What is Memwyre?
Memwyre is an AI memory platform. It keeps a shared memory across AI tools, agents, and conversations, and connects through MCP.
What Memwyre does
Memwyre compiles work from the tools into shared memory and recalls it before a tool acts. It is described as a context layer for models, editors, and terminal tools, not a standalone chatbot. IDE connections cover Cursor, VS Code, and Windsurf. Claude Code and OpenClaw are on the page for terminal agents. One memory layer is described across Claude, ChatGPT, and Gemini. A Chrome extension is listed. Install can use the MCP server, including the command npx -y install-memwyre. Code, notes, and memories are not used to train public models. Transit uses TLS 1.3 and data at rest uses AES-256-GCM.
Who Memwyre is for
Free is essential persistent memory for developers exploring Claude, Cursor, and ChatGPT. Pro adds connectors for people who want Drive and Notion sync. Enterprise is for a team that wants VPC self-hosting, SSO, and a shared vault. The use-cases page also lists software engineers, researchers and analysts, product managers, students and learners, and DevOps and automation engineers.
How Memwyre works
Sign up with email or GitHub. No credit card is required for Free. Connect Cursor, VS Code, or a terminal agent through MCP. The FAQ says connecting takes less than two minutes. Free includes 50 persistent memories and 10 document uploads. You can continue querying existing memories after that cap. Adding more, or turning on Notion and Google Drive sync, requires Pro. Cancel or downgrade keeps existing memories searchable, and new additions follow the Free caps.
Key features
- Persistent shared memory across browser, ChatGPT, Claude, and Cursor, searchable by the connected tools.
- MCP server, CLI session lifecycle hooks, and a Chrome extension.
- Document uploads on Free for PDF, MD, and code. The home FAQ also lists 30 AI chat messages on Free.
- Pro adds unlimited memories and documents, Google Drive and Notion sync, and plugins for Claude Code, OpenCode, and OpenClaw.
- Pro lists cross-encoder re-ranking (LoCoMo 94.6%).
- Enterprise lists self-hosted or VPC deployment, a shared memory vault, SAML 2.0 and Okta SSO, RBAC, zero data retention, a legal BAA, and a 99.9% uptime SLA.
- Encryption in transit is TLS 1.3. Encryption at rest is AES-256-GCM. Memories are not used to train public models.
Pricing
As listed on memwyre.tech, checked 14 October 2026. Prices can change; see the site for current terms.
- Free is $0 forever: 50 persistent memories, 10 document uploads (PDF, MD, code), the MCP server, CLI session hooks, and community Discord support. The home FAQ also includes 30 AI chat messages and says no credit card is required.
- Pro is $12/month: unlimited memories and documents, Google Drive and Notion sync, Claude Code, OpenCode, and OpenClaw plugins, cross-encoder re-ranking listed as LoCoMo 94.6%, and priority developer support.
- Enterprise is custom, via contact sales: self-hosted or VPC deployment, an organization shared memory vault, SAML 2.0 and Okta SSO, RBAC, zero data retention, a legal BAA, 99.9% uptime, and a dedicated Slack.
- The enterprise FAQ says volume billing is for engineering teams with 5 or more developers, or for SOC2, BAA, or Okta SSO needs.
Facts from memwyre.tech, checked 14 October 2026
- The LoCoMo benchmark includes Memwyre v2 and v1. It is not a price.
- Install text includes npx -y install-memwyre, or adding the remote MCP server to a configuration file.
- Setup is described as under 2 minutes, and also as less than two minutes.
Written from Memwyre's own website (Homepage, Pricing, Features). The maker can claim this listing to edit it.
Categories
Frequently asked questions
Is there a free plan?
Yes. Free includes 50 memories, 10 document uploads, and MCP. The home FAQ also lists 30 AI chat messages. No credit card is required. You can sign up with email or a GitHub account.
What happens after 50 memories?
You can continue querying and retrieving your existing memories. Adding more memories or documents, or enable automated syncing from Notion and Google Drive, requires Pro.
Can the plan be cancelled?
Yes. Upgrade, downgrade, or cancel from the account. Existing memories remain preserved and searchable. New additions then follow the Free caps.
Is the stored code private?
The FAQ says code, notes, and memories are never used to train public models. Data is encrypted in transit with TLS 1.3 and at rest with AES-256-GCM. Enterprise also offers self-hosted Docker and VPC deployment.
How does the LoCoMo figure on the benchmark differ from the Pro card?
The Pro card lists cross-encoder re-ranking as LoCoMo 94.6%. The LoCoMo benchmark reports non-adversarial accuracy of 70.5% on 1,540 questions, mean context of 4.9k tokens against 26k+ raw conversation tokens, and says v2 improves upon v1 by 24.4% average accuracy while reducing context footprint by 81%.