MENSARA.AI · A GOVERNED MIND, IN YOUR ACCOUNT

Decide if it fits · GUIDE 01

Decide whether Mensara fits a real job

Mensara helps people and connected assistants work from shared notes while keeping access tied to each person or assistant. It is useful when the same background comes up repeatedly and you want answers that lead back to a source you can inspect.

Picture one job

Suppose you are preparing for a customer conversation. You can ask an assistant to find the current account decision and the open questions in the notes it is allowed to read. Another assistant may see a different set of notes, even for the same customer. Access does not travel from one helper to another.

The useful result is not just a polished answer. It is an answer with a source you can open and judge.

When it is a good fit

Start with a recurring question grounded in records your team maintains: preparing a handoff, finding the current decision, or bringing a helper up to speed. Check whether the returned sources are the ones you would use yourself.

Mensara needs an owner who can decide what should be shared and review requests for broader access. It is less useful when the work has no maintainable sources, when nobody can own those boundaries, or when you expect every helper to see everything automatically.

An approval label records a status; it does not make the answer true. Check the source when the decision matters.

Related

Ask your first useful question · Where your data goes and who controls it

NEXT GUIDE · 02 See whether Mensara helps with your work Mensara is useful when people and assistants repeatedly need the same work context, while needing different access to it. →

MENSARA.AI · DECIDE WHETHER MENSARA FITS A REAL JOB · AGENT VIEW · TEXT/MARKDOWN

# Evaluation procedure

Mensara exposes a scoped knowledge repository through a gateway. Evaluate it for a shared-context read/proposal workflow, not as proof of general answer correctness. This release artifact is verified; a full MCP transport/client onboarding and deployed availability are not demonstrated.

Prerequisite: the human owner has completed the supported authenticated connection and granted your principal scope. Do not invent a signup or provisioning endpoint; autonomous provisioning is not documented.

Use discovery before retrieval. `orient` supplies the principal, scope, zones, first-read guidance, protocol, and memory-authoring information. Navigate returned zones and map files, then read linked sources. Zone IDs are not guaranteed to be directory paths. Cite the accessible source path and relevant passage in any answer.

**Template — not a connection instruction or runnable client transcript.**

```
orient({})
list_files({"dir":"<granted directory from discovery>"})
read({"path":"<discovered map path>","memory_view":"current"})
read({"path":"<authorized note linked by the map>","memory_view":"current"})
```

Parse each result. On denial, stop; do not substitute an owner identity. A successful read is expected to identify content, zone, and redactions, with memory metadata where relevant. It does not grant proposal authority and does not certify the answer's factual truth.

For a write candidate, discover the destination and policy, then use the applicable proposal flow. Report only its returned held/applied disposition. For an unprovisioned Mind, involve the human owner through the human-assisted onboarding route.

The public homepage currently offers a [founding-cohort request](https://www.mensara.ai/#founding-cohort) and a [Create account entry](https://www.mensara.ai/auth/signup). The latter reached an authentication page when checked; account creation, payment, provisioning and seat availability were not verified. Confirm those prerequisites before treating it as a self-service trial.

## Continue

[First agent read](/how-it-works/first-useful-question) · [Ownership and data flow](/how-it-works/ownership-data-flow)

[Generated tool input reference](/how-it-works/tool-inputs.md)