Question 1

What is MemQ?

Accepted Answer

MemQ is a hosted, MCP-native memory layer that gives AI agents persistent, governed memory. It lets Claude, Claude Code, Codex, Cursor, Gemini-style clients, OpenClaw, Legion, and any MCP-compatible client recall context across sessions, share approved knowledge between agents through a governed commons, and hand work forward without losing state. MemQ is more than vector search: it adds journaling, reflection, sleep-style consolidation, checkpoints, and exact replay.

Question 2

How is MemQ different from Mem0?

Accepted Answer

Mem0 is an SDK you wire into each application; MemQ is a hosted MCP server your agents connect to directly — no SDK, no lock-in. MemQ adds a governed commons layer so one agent's approved learning becomes shared context for every agent, plus sovereign namespace isolation, OAuth and API-key access, and reflection and consolidation. If your client speaks Model Context Protocol, it speaks MemQ.

Question 3

Does MemQ work with Claude Code, Cursor, and Codex?

Accepted Answer

Yes. MemQ works anywhere MCP is accepted. It ships verified support for Claude Code and Codex, config packs for Cursor, VS Code, Windsurf, and Antigravity, and connects to any MCP-compatible IDE, CLI, or agent runtime through a single hosted endpoint at https://mcp.multinex.ai/mcp/v1.

Question 4

How much does MemQ cost?

Accepted Answer

MemQ starts free with no credit card. The free tier includes a private starter namespace with 100 memory segments, 25 writes, 150 reads, and 5 AI calls. Paid plans scale up: Contributor at $20/month and Pro at $50/month for individuals, plus Team plans for shared organization namespaces. Authenticated @multinex.ai users access MemQ at no charge.

Question 5

Is my memory private and secure with MemQ?

Accepted Answer

Yes. MemQ is sovereign by design. Your memory lives in an isolated namespace with zero cross-tenant leakage by default, private memory never becomes shared knowledge without explicit promotion controls, every access is governed and auditable, and you can export your data at any time. A self-hosted option is available.

Question 6

What is the MemQ commons?

Accepted Answer

The MemQ commons is a governed, shared memory layer across your agents. When one agent learns something approved for sharing, every authorized agent can recall it — namespace-scoped, attributed, and separated from private customer memory by default. It turns approved external intelligence and repeated internal lessons into reusable agent context.

Question 7

Is MemQ just a vector database?

Accepted Answer

No. Vector search is one retrieval path among several. MemQ fuses recent context, semantic recall, journal evidence, and exact replay, then adds a Soul Journal for decisions and failures, reflection and sleep-style consolidation for durable learning, checkpoints for resumable plans, and namespace-aware governance — capabilities a vector store alone does not provide.

Shared memory for every agent you run.

Works anywhere MCP is accepted.

One agent learns it. Every agent knows it.

Fast recall. Open methodology.

Your memory. Your namespace. Your rules.

Connect once. Remember forever.

Connect

Work

Return

MemQ is the foundation. Build on it.

Shield

Legion

AI Transformation

Start building shared memory in 30 seconds.