Docs
Find one useful job for your Mind.
Start with the guide that matches the work in front of you. When an agent needs exact steps, switch the page to Machine or open its Markdown version.
Content updated .
Operating procedures
- Connect a coding client and verify identity: Use the Mind’s trusted connection guide and the person’s OAuth sign-in, then verify the actual Mind tools.
- Find an authorised answer and verify its sources: Discover the intended Mind, follow its maps, open the sources and report only the answer they support.
- Read, correct and verify memory: Find the authorised source, submit one bounded correction and report the actual proposal state and current wording.
- Invite a member and verify effective access: Issue one authorised invitation, then distinguish invitation, accepted membership and the recipient's effective access.
- Build, propose and verify a governed model agent: Construct and test a real model-agent bundle, submit it for review, install separately through an authorised interface, and observe its intended scheduled output.
- Make Mind calls, apps and agents fast: Exact speed rules for a connected agent and for an installed app built with the app-engineer skill: batched reads and writes, concurrent model calls, backfill runs and the sandbox time limit.
- Compose and supervise a bounded loop: Propose a loop, inspect its state, and hand approval or activation to an authorised supported interface. Verify a case before reporting execution.
- Inspect and change effective teammate access: Use the advertised profile tools for an authorised change, then check the result with the intended person’s own account.
- Use Portal chat and verify its effects: Follow the normal signed-in Portal route; do not call internal chat APIs as if they were public agent tools.
- Connect and inspect an external source: Discover the actual interface, let the owner complete service consent and verify the connection separately from app execution and saved capture.
- Capture information and verify the saved result: Keep the source and date, choose an authorised destination, submit one bounded proposal and report applied or held accurately.
- Read current memory and historical evidence: Choose the view that fits the requested claim and retain the distinction between current guidance and earlier evidence.
- Inspect the Inbox and route each human decision: Read the requests visible to the caller, identify their kind and use the supported action for that specific object.
- Create, update and verify a task: Use the canonical task tools to keep one saved work record with a checkable outcome. Discover the intended Mind, project and authorisation before writing.
- Renew a task claim and hand work over: Keep durable task progress separate from the short-lived runtime session that owns updates. Use the session and claim tools the intended connection actually advertises.
- Draft and verify a decision or policy record: Find the saved evidence for a settled decision or rule, propose a structurally valid operating record and verify its actual saved status.
- Discover and open a published Mind View: Use the active member's own connection to discover a published View, resolve its address and inspect the authorised data result. View tools require a Mind version with the released View feature.
- Inspect the governance Inbox and answer the right item: Read the caller-scoped attention worklist, inspect each original request and report only the outcome actually observed.
- Discover a skill and prepare a tested revision: Use the Mind's current skill discovery to select a procedure the connected client can perform. Treat authoring and executing a skill as separate work.
- Compose, approve and inspect an automation: Build a trigger and ordered action chain from the live installed palette. Approval can enable future dispatch, so agree the intended effects before the lifecycle action.
- Read a loop case and reconcile a wait: Use the returned loop and case identifiers to inspect real recorded progress. Loop management reads require the relevant operate_loops authority; being able to read ordinary company notes is not that grant.
- Prepare a loop correction and replay its replacement: Turn recorded corrections into a proposed replacement. Keep composition, replay, approval and activation as separate observed stages.
- Inspect and change installed-agent controls: Use the intended app identity and current lifecycle authority. Model budget, app access, grant expiry, trigger and loop autonomy are separate controls.
- Install a selected sensor and verify capture: Check the exact selected bundle, required connection and normal scheduled output without treating store selection or a liveness probe as capture acceptance.
- Validate, approve and publish a View: Build a complete read-only View bundle, obtain independent approval and verify its published route and data as an intended viewer.
- Connect and share through a Mind’s Network: Use the owner-controlled peer lifecycle, inspect installed action authority and report transport and saved-content states separately.
- Inspect model routes and spending controls: Distinguish personal chat routes, app model consent, rolling app budgets and owner route-sharing controls before changing any of them.
- Inspect account, payment and Mind readiness: Use the normal human account interface to distinguish organisation membership, billing state and a ready workspace without exposing an account-management API as an agent tool.
- Connect a backup and inspect recovery readiness: Use the owner-authorised backup ceremony, inspect freshness and kit state, and hand sensitive recovery material through the protected owner route.
- Diagnose workspace availability without changing work: Identify the failing boundary, inspect read-only state and preserve the original action reference before considering a retry.
- Operate the protected Linux Network companion: Install a supported participant, separate service/human/agent authority and complete exact-message approval before transport.
Store and share knowledge
Keep records with their history and decide who may use each area.
- 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.
- How zones and permissions work
A zone is an area of your Mensara Mind with its own access rules.
- Team and roles
A team shares a Mensara Mind through separate accounts.
Use your Mind
Choose a connection and understand how saved information supports work.
- Portal, MCP and CLI
The Portal is your Mind’s web interface.
- Bring email, calendar events and files into your Mind
A connection lets an installed app use an external service you have authorised.
- Understand skills, agents, automations and loops
Your Mind can hold instructions for your AI tool and run installed programs with agreed access.
Keep control
Review changes, understand settings and prepare for recovery.
- Review changes and check what was saved
A proposal is a suggested change to information saved in your Mensara Mind.
- Understand your Mind’s settings
Your Mind is the Mensara workspace holding your team’s saved information and work.
- Back up, recover and export your Mind
A backup is a second copy kept away from the server running your Mind.
- Set up your Mensara Mind and get a first useful answer
Your Mind is the Mensara workspace that keeps your team's saved information and work.
Before you start
Find an answer to a problem or check the conditions before relying on a feature.
- Find help by what went wrong
Start with the action you were trying to complete and the result you actually saw.
Decide if it fits
Start with a real job and a result you can check.
- Decide whether Mensara fits a real job
Mensara gives your team a shared Mind: saved information and work that people and their connected AI tools can use.
- See whether Mensara helps with your work
Mensara is useful when people and their AI tools keep needing the same background, while each person should see only the information their work requires.
Get a first result
Ask a small question, inspect its source and choose the right chat route.
- Start with a question you can check
A good first Mensara job starts with one question and a source you can inspect.
- Ask your first question
Your Mind is the Mensara workspace holding your team’s saved information.
- Use chat in your Mensara Portal
Chat in your Portal lets you ask an AI model questions and work with your Mind’s records in the same web interface.
Use memory well
Save useful information, keep it current and recover from a poor answer.
- Start from what’s already decided
Saved context is the information a later conversation can use without you rebuilding the whole briefing.
- Ask what is current, then ask why
Current information tells you what should guide a job now.
- Ask to see the source behind an answer
A map in your Mensara Mind is a guide to related records.
- Save a note, meeting or useful fact
Saving information in your Mensara Mind gives a later person or agent a record to find and check.
- Recover from a missing or poor answer
An AI answer can be poor because the source is missing, outdated, outside your access or misunderstood.
- Understand activity and the history of a change
Activity records help explain which identity asked for something and what the Mind did.
Set safe boundaries
Choose access for the job and check sharing or removal as the intended person.
- Choose access for a teammate
Access is the information and actions a person is allowed to use in your Mensara Mind.
- Know what an agent may read or change
Reading information, suggesting a change and deciding a request are different actions in Mensara.
- Decide where your information may go
Your Mensara Mind holds your team’s saved information and work.
- Correct saved information in your Mind
Mensara memory is the information your team saves in its Mind for later use: decisions, notes and ways of working.
- Invite a teammate to your Mind
A Mensara Mind holds your team's shared information and work.
- Join your team’s Mind
Your team’s Mind is its shared collection of information and work in Mensara.
- Share work without sharing everything
Sharing work in Mensara means giving the intended people a useful source and the access needed for their task.
- Check what your team promised a customer
Work through a fictional handoff and distinguish an agreement, a request and a pending record edit.
- Change or remove a teammate’s access
A teammate’s effective access is what they can actually read and do after roles, profiles, defaults and other grants are considered.
Manage work and decisions
Track progress, preserve decisions and open useful team Views.
- Track tasks and check their progress
A task is a saved record of work somebody has agreed to do.
- Help another agent continue the work
A handoff gives the next person or agent enough saved context to continue a specific job.
- Record a decision or policy with its evidence
A decision record explains a choice your team has already made and why it applies.
- Review and replace an operating record
An operating record is a saved decision or policy with its owner, scope, evidence and review date.
- Open a team View and check its information
A View is a small web page on your Mind that presents saved records as a board, table, list or dashboard.
- Review what needs your attention
Your Mind's Inbox brings together requests that need a human answer or decision: held changes, loop questions, overdue records and work needing attention.
Run agents, automations and loops
Build a bounded job, inspect its cases and control recurring work.
- Choose the right kind of help
Mensara gives you several ways to work with the saved information in your Mind.
- Use a skill or turn a repeated job into one
A skill is a written procedure your connected AI tool can read and follow.
- Install an approved agent and check its first result
An installed agent is a program that does a defined job on your Mind with its own permissions.
- Build a governed AI agent
Your Mind is your team's shared Mensara workspace for saved information and work.
- Make apps and agents on your Mind fast
Speed on a Mensara Mind comes mostly from making fewer, larger calls and keeping each run short.
- Pause, upgrade or remove an installed agent
An installed agent is a program running a defined job on your Mind with its own access, connections and model limits.
- Automate a routine task with installed actions
An automation starts an ordered sequence of actions when its trigger occurs.
- Build your first recurring workflow with a loop
Your Mind is the Mensara workspace holding your team's saved information and work.
- Understand a loop case and follow its result
A loop is a recurring workflow with steps, checks, branches and possible waits.
- Answer a loop exception and correct repeated mistakes
A loop exception is a case that needs information, a person's decision or attention after a failed check.
- Control agent autonomy, budgets and access
Autonomy is how much a workflow may do without stopping for a person.
Manage your workspace
Connect models, understand spending and inspect account or workspace health.
- Manage models and service connections
A model generates an AI response.
- Understand usage, budgets and costs
Usage records what was consumed.
- Manage your Mensara account and billing
Your Mensara account holds your organisation membership, plan and Mind address.
- Manage notifications and check your Mind’s health
Notifications bring selected work to your attention.
Get help and check changes
Reconnect, reconcile interrupted work and check which instructions apply.
- Fix sign-in and reconnect the right service
Your Mensara account, your Mind’s Portal, an AI client and a model provider can each have their own session.
- Recover pending, failed or interrupted work
When a connection stops, the last action may have failed, may still be running or may already have completed without its response reaching you.
- Check your version and understand product changes
A released feature is available only when the required software, configuration, permissions and installed app are present in your Mind.
Browse and connect
Check selected extensions and their source, connection and review requirements.
- Choose an app, sensor or connector for your Mind
An installed app does a defined job inside your Mind.
- Capture Telegram chats in your Mind
Telegram Sensor copies messages visible to a dedicated Telegram bot into your Mind’s inbox.
- Plan SharePoint capture and export for your Mind
SharePoint capture brings documents from an approved library or folder into your Mind.
- Review records imported from business tools
An imported record is information copied or interpreted from another system, such as a customer note or contact update.
Build a team View
Turn declared Mind datasets into a reviewed page with checked access.
- Build and publish a team View
A View is a small web page on your Mensara Mind that presents saved records as a board, table or dashboard.
Connect and share
Share a specific piece of work between separate identities and inspect delivery.
- Connect Minds with Mensara Network
Mensara Network lets an authorised owner connect their Mind to another participant and share selected work.
- Share work and reply across Minds
A Network Share sends selected text to an authorised peer.
- Connect an external agent to Mensara Network
An external Network participant can exchange approved Shares and Replies without running a full Mensara Mind.
Standalone client
Set up the supported Linux companion and keep human approval separate from the agent.
- Set up the Linux Network companion
The Network companion lets a restricted external agent prepare messages through a separate trusted Linux service.
- Prepare, approve and follow an external Network message
The Linux companion separates a draft from permission to send it.
- Recover and migrate a Network participant
The external companion keeps identity, pending approved work and history in protected local state.
Interface reference
Discover the installed interface, choose a procedure and verify the actual result.
- Agent interface reference
Use this reference to identify the right interface, understand its effects and choose a procedure you can complete with the connected client.
Agent references
Exact inputs, when you need them. These references list each tool's exact inputs and mark which behaviour has been tested.