MemoryCrow
Sign inGet started for free

FAQ

Questions teams ask before they trust shared memory.

Short answers on setup, pricing, permissions, source handling, MCP, and how MemoryCrow fits around the tools your team already uses.

Getting started

How MemoryCrow fits around the tools and AI surfaces your team already uses.

Does MemoryCrow replace Slack, Notion, GitHub, Linear, or our help desk?

No. Those tools stay where work happens. MemoryCrow is the memory layer beside them: save or import the context worth remembering, then recall it from the AI tools your team already uses.

See integrations

What is the smallest useful setup?

Start with one useful source of truth: a decision, policy, support answer, PR note, or customer handoff. Add it to MemoryCrow, then ask for it from Claude, Cursor, ChatGPT, or another MCP client.

Do we need a migration project before trying it?

No. Start with the context that keeps getting asked for. You can add more memories, imports, and integrations over time without moving the team into a new wiki.

Which teams get value first?

Support, sales, and engineering usually feel the pain first because their answers change hands often: policies, customer promises, rollout notes, architecture decisions, and incident history.

Browse use cases

Memory and recall

The difference between what you store, what the AI can ask, and what comes back.

What counts as a memory?

A memory is durable company context saved for later recall: a decision, note, ticket, doc section, handoff, policy, or source-backed fact. It keeps the source trail attached.

View pricing

Is recall unlimited?

Yes. Plans meter stored memory, not the number of questions your team asks. Claude, Cursor, ChatGPT, and MCP-compatible clients can keep asking against the same memory layer.

What happens when two sources disagree?

The newer or approved source can supersede the stale one without deleting history. The old fact stays traceable, but it stops becoming the answer.

What happens if MemoryCrow cannot prove an answer?

It abstains. If there is no relevant stored memory, no current fact, or no allowed scope for the asker, MemoryCrow should return nothing instead of inventing a confident answer.

Permissions and safety

How MemoryCrow decides who can see an answer and why some answers stay hidden.

How does permission-safe recall work?

Memory is stored with a scope: personal, workspace, or group. Recall filters by that scope before returning facts or citations, so private memory does not become a team answer by accident.

Permission-safe recall

Can two people ask the same question and get different answers?

Yes. If one person has access to a personal or group-scoped memory and another does not, MemoryCrow can return a cited result for one person and abstain for the other.

Does connecting or importing a source make everything visible to everyone?

No. Capture and import flows use the scope you choose. Connecting Slack or Notion does not automatically make all of that workspace visible to every MemoryCrow user.

Can we review data processing and subprocessors?

Yes. The legal and trust pages explain privacy, data processing terms, and the third parties MemoryCrow uses to process data.

DPA Sub-processors

AI tools and MCP

How Claude, ChatGPT, Cursor, VS Code, and MCP clients use the same memory layer.

What is MCP in plain terms?

MCP is a standard way for AI tools to call external tools. MemoryCrow exposes your saved memory over MCP so those tools can retrieve source-backed context.

MCP server

Which AI clients can connect?

Claude, Cursor, ChatGPT, VS Code, and other MCP-compatible clients can connect to MemoryCrow.

Can agents write to memory?

Yes. MCP clients can save useful facts, decisions, and handoffs into MemoryCrow, using the scope selected for that memory.

Can access be revoked later?

Yes. MCP authorization can be revoked, and MemoryCrow scopes can be changed or removed when access should stop.

Pricing and billing

Plans, memory limits, trials, and when one-time packs or Enterprise make sense.

What is the main difference between Personal and Team?

Personal is for one person. Team is a flat workspace plan with more included memory and members. The memory layer itself stays the same: cited recall, source trails, current facts, and permission scope.

Compare plans

Can we buy more memory without changing plans?

Yes. One-time memory packs can be added for imports, launches, migrations, or large backfills. Packs stack on top of the workspace plan.

Do you charge per seat?

No. Team is a flat workspace plan up to the listed member limit, so the bill does not grow every time another teammate needs cited recall.

When should we talk Enterprise?

Use Enterprise when procurement, SSO, security review, private deployment paths, or custom memory limits matter before rollout.

Book a demo

Data and model training

What happens to your company memory and what MemoryCrow does not claim.

Do you train models on our data?

No. MemoryCrow stores and recalls your workspace context. It does not train foundation models on your company data.

Where can we read the privacy policy?

The privacy policy explains how MemoryCrow collects, uses, protects, and gives you control over your data.

Privacy policy

What if we need to cancel or request a refund?

The refund policy explains cancellations, billing changes, and refund handling for MemoryCrow plans.

Refund policy

Who should we contact for a security or data question?

Email [email protected]. You can also use the contact page if you want the request routed from the website.

Contact

Start with one memory

Give every AI the same trusted memory.

Start the trial, connect one source, and ask the next question from Claude, ChatGPT, Cursor, or any MCP client. If MemoryCrow cannot prove it, it will not guess.

Start free trial Book a demo
  • 7-day trial
  • Unlimited recall
  • Permission-safe by default