What is ContextVault?

A shared memory layer for AI systems. Store a useful memory once, and every AI client your team uses can recall it, reuse it, and build on it.

Does it work with multiple models and providers?

Yes. Anything that speaks MCP plugs in, including Claude, Copilot, ChatGPT, and Ollama. The Vault is model-agnostic by design, your team can use whatever client they want.

How is memory scoped?

Every entry is scoped to an organization, and gated further by role and group ACLs. Nothing crosses tenant boundaries.

How long does integration take?

Seconds. Add the MCP server to your AI client, and ContextVault is immediately available to use.

Is there a free plan?

Yes. The free trial gives one user a single default group and 50 memories to test the workflow.

How do security and tenant isolation work?

Shared Postgres with org_id enforcement in the app layer, encryption at rest, and per-tenant audit logs. Dedicated single-tenant deploys are available on Enterprise.

Can I migrate from another vector store?

Not yet. Importing existing memories is planned as a soft beta feature after the core store and search workflow is stable.

Can I use ContextVault on multiple machines?

Yes, if you have an account you can add the MCP to any machine you have with AI and use it. There are no restrictions other than you must have a valid, authenticated account.