Store and share knowledge

How Mensara memory works

Memory in Mensara is the information your team saves so people and their agents can use it later: decisions, notes and ways of working. Your Mind keeps those records together with their sources and history. This page explains how information is saved, found and corrected, so you can judge the answers you get.

Contents

What memory means here

Think of the Mind as your team's shared written reference. A record is one saved piece of information, such as a decision or a customer note. Your connected agent is the tool you ask to find, explain or suggest changes to those records. The answer it writes in your chat is different from the underlying saved information.

For example, a teammate can ask what the current support hours are. Their agent finds the relevant saved record, uses it to answer and names the source so the teammate can check it. The agent can only use information that account is allowed to read. A wrong or missing source can still produce a poor answer, so ask to see the evidence before relying on it.

An answer uses saved information

The record is kept in your Mind. The answer appears in your chat.

Your Mind holds saved records with their sources and history. When you ask a question, your connected agent reads relevant records your account is allowed to access. It uses those records to write an answer and names the sources so you can open them. The chat answer is separate from the saved record. Writing or correcting an answer in chat does not by itself show that a record was saved or changed.

Where information is kept

Your Mind stores readable text files. They use Markdown, a plain-text format for headings, paragraphs and links, and Git, a version-history system that keeps earlier saved versions. That history is private to your Mind and is carried in its backup; it is not published anywhere. Folders give it a visible shape: one area might hold customer notes, another company policies, another the decisions from leadership meetings.

Areas have a short map file that says what is in them and where to look. People and agents start from those maps. A map is a starting point, not a full list.

Review new information

An email, a meeting note or an update from a teammate's agent can arrive as a suggested change. Items from a connected service carry their source; an agent's suggestion carries who sent it and why.

Whether a change waits for a person or is saved straight away depends on the permissions and write rules for that area. By default, a change is saved without a person only when an automatic check passes and both the proposer's trust level and the area's own limit allow it. Deleting or renaming a file always waits for the owner or another approver, and areas that hold governed material, such as decisions, typically hold changes from members for the owner. Someone can save a held change, correct it or leave it out. Saved changes keep their earlier versions in the history. Understand the settings that govern review.

Current information and its history

Not every record should guide today's work. In your Mind's memory areas, each entry carries its own status, and that status decides whether it counts as current. An entry can be current, waiting for approval, superseded (replaced by a newer one), expired, or in conflict with another. "Waiting for approval" here means an entry that has been saved but not yet approved. It is different from a proposed change held for review, which has not been saved at all.

Every memory entry also has a time limit. The starter template sets 30 days, counted from the date the entry was last verified or, if it has none, created. When the limit passes, the entry stops appearing in current answers until someone verifies or refreshes it. Foundational entries can be set never to expire. Ordinary notes outside the memory areas have no status of this kind, so check their own wording and date.

Ask for current guidance when deciding what to do now, and for history when you want to understand a change. The current and historical views help distinguish them. Check dates and sources when records disagree; an answer is not proof that the underlying information is correct.

The current record and its earlier version

History explains a change. Check the record’s status and sources before using it today.

In this fictional example, an earlier saved instruction says to use the blue folder. A reviewed correction changes the saved instruction to use the green folder. The earlier version remains in history so the change can be understood. A later date alone does not make guidance current: check its status and evidence, including whether it is superseded, expired or in conflict.

Finding an answer

Ask a question such as "What is our current refund policy?" Your agent searches only the areas your account may read.

Ask it to name the files it used, so you can open the source before you rely on the answer.

Correct information that is wrong

Ask your agent to open the saved source, explain the correction and propose the new wording with its evidence. A proposal is a suggested edit. It may be saved under the Mind's rules or wait for somebody with approval authority. After it is saved, open the record again to check the result. Correct a record and check what changed walks through that job.

Related

Practical guides for people and agents

Start with one useful job

Join the waitlist

Plain text for AI agents. Same content as the human page.

## For agents

- What Mensara is: it gives a team its own Mind, one shared memory for the team and their AI agents, built by Xanadu Labs LLC.
- Status: signup is not open. A person joins the waitlist at https://www.mensara.ai/#waitlist. Agents cannot create an account or sign up.
- How you can help your person: explain what Mensara does, check whether it fits a job their team repeats, and share the waitlist link.
- Agent references: https://www.mensara.ai/llms.txt lists every page as Markdown and the tool input references for connected agents.

---

# How Mensara memory works

Content updated 29 October 2026.

How Mensara works · Store and share knowledge · [How it works](https://www.mensara.ai/how-it-works)

Memory in Mensara is the information your team saves so people and their agents can use it later: decisions, notes and ways of working. Your Mind keeps those records together with their sources and history. This page explains how information is saved, found and corrected, so you can judge the answers you get.

[Page as Markdown](https://www.mensara.ai/how-it-works/memory.md)

## What memory means here

Think of the Mind as your team's shared written reference. A **record** is one saved piece of information, such as a decision or a customer note. Your connected agent is the tool you ask to find, explain or suggest changes to those records. The answer it writes in your chat is different from the underlying saved information.

For example, a teammate can ask what the current support hours are. Their agent finds the relevant saved record, uses it to answer and names the source so the teammate can check it. The agent can only use information that account is allowed to read. A wrong or missing source can still produce a poor answer, so ask to see the evidence before relying on it.

### An answer uses saved information

The record is kept in your Mind. The answer appears in your chat.

Your Mind holds saved records with their sources and history. When you ask a question, your connected agent reads relevant records your account is allowed to access. It uses those records to write an answer and names the sources so you can open them. The chat answer is separate from the saved record. Writing or correcting an answer in chat does not by itself show that a record was saved or changed.

## Where information is kept

Your Mind stores readable text files. They use **Markdown**, a plain-text format for headings, paragraphs and links, and **Git**, a version-history system that keeps earlier saved versions. That history is private to your Mind and is carried in its backup; it is not published anywhere. Folders give it a visible shape: one area might hold customer notes, another company policies, another the decisions from leadership meetings.

Areas have a short map file that says what is in them and where to look. People and agents start from those maps. A map is a starting point, not a full list.

## Review new information

An email, a meeting note or an update from a teammate's agent can arrive as a suggested change. Items from a connected service carry their source; an agent's suggestion carries who sent it and why.

Whether a change waits for a person or is saved straight away depends on the permissions and write rules for that area. By default, a change is saved without a person only when an automatic check passes and both the proposer's trust level and the area's own limit allow it. Deleting or renaming a file always waits for the owner or another approver, and areas that hold governed material, such as decisions, typically hold changes from members for the owner. Someone can save a held change, correct it or leave it out. Saved changes keep their earlier versions in the history. [Understand the settings that govern review](https://www.mensara.ai/how-it-works/mind-settings#why-changes-wait-for-review).

## Current information and its history

Not every record should guide today's work. In your Mind's memory areas, each entry carries its own status, and that status decides whether it counts as current. An entry can be current, waiting for approval, superseded (replaced by a newer one), expired, or in conflict with another. "Waiting for approval" here means an entry that has been saved but not yet approved. It is different from a proposed change held for review, which has not been saved at all.

Every memory entry also has a time limit. The starter template sets 30 days, counted from the date the entry was last verified or, if it has none, created. When the limit passes, the entry stops appearing in current answers until someone verifies or refreshes it. Foundational entries can be set never to expire. Ordinary notes outside the memory areas have no status of this kind, so check their own wording and date.

Ask for current guidance when deciding what to do now, and for history when you want to understand a change. The current and historical views help distinguish them. Check dates and sources when records disagree; an answer is not proof that the underlying information is correct.

### The current record and its earlier version

History explains a change. Check the record’s status and sources before using it today.

In this fictional example, an earlier saved instruction says to use the blue folder. A reviewed correction changes the saved instruction to use the green folder. The earlier version remains in history so the change can be understood. A later date alone does not make guidance current: check its status and evidence, including whether it is superseded, expired or in conflict.

## Finding an answer

Ask a question such as "What is our current refund policy?" Your agent searches only the areas your account may read.

Ask it to name the files it used, so you can open the source before you rely on the answer.

## Correct information that is wrong

Ask your agent to open the saved source, explain the correction and propose the new wording with its evidence. A proposal is a suggested edit. It may be saved under the Mind's rules or wait for somebody with approval authority. After it is saved, open the record again to check the result. [Correct a record and check what changed](https://www.mensara.ai/how-it-works/suggest-a-correction) walks through that job.

## Example: A sales call becomes part of your Mind

In this illustrative example, after a call with a customer, a salesperson's agent suggests a note for that customer's files and says where it came from. After the note is saved, a person with access can retrieve it and inspect its source when asking about the account.

## Related

- [Correct saved information in your Mind](https://www.mensara.ai/how-it-works/suggest-a-correction)
- [Start from what’s already decided](https://www.mensara.ai/how-it-works/carry-context)
- [Ask what is current, then ask why](https://www.mensara.ai/how-it-works/current-and-evidence)
- [Ask to see the source behind an answer](https://www.mensara.ai/how-it-works/maps-and-navigation)

Next page: [How zones and permissions work](https://www.mensara.ai/how-it-works/zones-and-permissions)

A zone is an area of your Mensara Mind with its own access rules.

[Join the waitlist](https://www.mensara.ai/?guide_attribution=eyJfcmFpbHMiOnsiZGF0YSI6eyJvcmlnaW5fcGF0aCI6Ii9ob3ctaXQtd29ya3MvbWVtb3J5IiwiY3RhX2lkIjoibWFya2Rvd24iLCJjb250ZW50X3JldmlzaW9uIjoiOGI2ZGFjYTYxZDU5N2RmNCJ9LCJwdXIiOiJndWlkZS1jbGljayJ9fQ%3D%3D--c4a5954670010ae82908cd64e6789fb07179f8b2#waitlist)