Why cMCP
Memory with governance, not just memory
Most approaches give an AI some memory. Few give a company governed flow: who may write, who may read, and who decides what becomes shared knowledge.
Comparison
Where cMCP sits among the alternatives
We compare categories of product, not named products. Each category has strong options for the job it was built for.
| Capability | Developer memory kits (e.g. Mem0, Zep, Letta) | Vendor-built assistant memory | Enterprise search | Governed context layers | cMCP |
|---|---|---|---|---|---|
| Works with any AI app via open standards | Partly | No | Partly | Partly | Yes (most vendors in testing) |
| Learns new facts from AI work | Yes | Yes | No | Yes | Yes |
| Scopes person, team, project and company, with approval | Partly | Partly | No | Yes | Yes |
| Client walls | Partly | No | Partly | Yes | Yes |
| Runs on your own servers | Partly | No | Partly | Partly | Yes |
| Audit of every read, write and approval | No | Partly | Partly | Yes | Yes |
| Mature at large scale | Partly | Yes | Yes | Partly | Not yet (pilot) |
Categories vary by product and change quickly; this is our general view as of October 2026. Check the specifics of any product you are evaluating.
Our position
What we are, and what we are not yet
Where cMCP is strong
- One memory shared by people and many different AI apps, over MCP and REST; change vendor or use several at once (multi-vendor access is in testing)
- Promotion, inheritance and synthesis with explicit approvals
- Client walls and audience labels for keeping material apart
- An audit trail of what the AI knew
- Runs in your cloud or on your premises
Where we are honest about limits
- cMCP is in a live pilot. It is not yet mature at large scale, and large-scale load testing is on the roadmap.
- We cover a deliberately narrow set of read-only connectors today.
- It is not a fit for very small teams or one-off work.
Your people are not forced onto one brand of laptop or email client. Why should your AI memory be locked to one model?
See how cMCP could work in your organisation
We are onboarding a small number of design partners to the pilot programme. Tell us about your teams and the AI tools they use, and we will scope a pilot together.