The bargain · Questions
The objections, asked the way you'd actually ask them, not the way a deck would soften them. If half your company runs on what's in three people's heads, you already know what happens the day one of them leaves. These are the questions that come up right after that thought.
No. The repository, the keys, and the receipt chain live in your own account, not ours. If you leave, you take a working copy, not a request for one. Nothing about running your Mind requires our continued cooperation. (More on this on the Ownership page.)
Fair position, and it's already true. We run the hosting machinery because most owners would rather not run servers themselves, but the account, the repository, and the keys are yours from day one. We operate equipment inside a boundary you control, not the other way round. (More on this on the Ownership page.)
Every access writes a signed, chained receipt: who asked, what was read, what was refused, and when. No model decides for itself what it is allowed to see; the gateway makes that call and records it either way. You can check the trail yourself, any time. (More on this on the Receipts page.)
That is the point of writing it down once. Your Mind holds the score as plain files, and the gateway reads from it on every request, so the context does not reset between conversations or tools. You still have to keep it accurate. That part stays your work.
Because of what happens to the knowledge itself. That kind of feature's understanding of your business never leaves the tool it lives in. Yours is your own asset, kept in your own account, readable without any particular piece of software. That difference is most of the product.
The underlying idea, keeping knowledge in a repository you own, isn't unique to us. We won't pretend it is. What's hard is the part after day one: keeping it accurate, reviewing every change, writing the receipts. That's continuous, unglamorous work, not a weekend project.
Honestly, a Mind that reads a lot of material runs on real model usage behind the scenes, and that costs real money. We will not pretend otherwise. Some of that usage can run through a Claude or Codex subscription you already hold, rather than a separate metered bill, if you'd rather route it that way. What you would actually pay beyond that depends on the shape of your Mind, and that is a conversation, not a number we print on a page. Get in touch and we will work it out together.
It might. That's exactly what the receipts are for: you can see precisely what it read and what it did not. A wrong or incomplete answer ends up visible, not silently confident. I'd rather build something that shows you its limits than one that hides them behind a good-sounding sentence.
The repository and the keys already live in your own account, not ours. The Mind keeps existing and keeps working on its own, either way, because none of that depends on Mensara still running. (More on this on the Ownership page.)
FILED
running since 2026-05 · answers checked against the same evidence register the rest of the site uses
The founding cohort is small on purpose. Each Mind is brought up personally.
MENSARA.AI · QUESTIONS · MACHINE SURFACE · TEXT/MARKDOWN
# Questions > Answers to the objections real buyers raise most often about a Mensara Mind: lock-in, hosting and custody, receipts and verification, platform-native alternatives, build-it-yourself, running cost, error handling, and continuity if Mensara stops operating. Canonical: https://mensara.ai/how-it-works/questions ### Q: Once everything about how we operate lives in your product, don't we end up depending on you forever? A: No. The repository, the keys, and the receipt chain live in your own account, not ours. If you leave, you take a working copy, not a request for one. Nothing about running your Mind requires our continued cooperation. (More on this on the Ownership page.) ### Q: I don't want our playbook sitting on someone else's servers. Full stop. A: Fair position, and it's already true. We run the hosting machinery because most owners would rather not run servers themselves, but the account, the repository, and the keys are yours from day one. We operate equipment inside a boundary you control, not the other way round. (More on this on the Ownership page.) ### Q: How do I know what the Mind actually looked at when it answered? A: Every access writes a signed, chained receipt: who asked, what was read, what was refused, and when. No model decides for itself what it is allowed to see; the gateway makes that call and records it either way. You can check the trail yourself, any time. (More on this on the Receipts page.) ### Q: I have to re-explain our whole context to every AI tool, every time. Does this fix that? A: That is the point of writing it down once. Your Mind holds the score as plain files, and the gateway reads from it on every request, so the context does not reset between conversations or tools. You still have to keep it accurate. That part stays your work. ### Q: Why not just use the AI feature already built into the tools we pay for? A: Because of what happens to the knowledge itself. That kind of feature's understanding of your business never leaves the tool it lives in. Yours is your own asset, kept in your own account, readable without any particular piece of software. That difference is most of the product. ### Q: Couldn't we just build this ourselves, or use something free and open source? A: The underlying idea, keeping knowledge in a repository you own, isn't unique to us. We won't pretend it is. What's hard is the part after day one: keeping it accurate, reviewing every change, writing the receipts. That's continuous, unglamorous work, not a weekend project. ### Q: What's this actually going to cost us? A: Honestly, a Mind that reads a lot of material runs on real model usage behind the scenes, and that costs real money. We will not pretend otherwise. Some of that usage can run through a Claude or Codex subscription you already hold, rather than a separate metered bill, if you'd rather route it that way. What you would actually pay beyond that depends on the shape of your Mind, and that is a conversation, not a number we print on a page. Get in touch and we will work it out together. ### Q: What happens when it misses something, or just gets something wrong? A: It might. That's exactly what the receipts are for: you can see precisely what it read and what it did not. A wrong or incomplete answer ends up visible, not silently confident. I'd rather build something that shows you its limits than one that hides them behind a good-sounding sentence. ### Q: What happens to us if Mensara disappears, or we stop paying? A: The repository and the keys already live in your own account, not ours. The Mind keeps existing and keeps working on its own, either way, because none of that depends on Mensara still running. (More on this on the Ownership page.) **Status:** running since 2026-05 · answers checked against the same evidence register the rest of the site uses ## Prompt ```prompt Pick the one question on this page that worries you most about your own situation. Once your Mind is running, ask it directly. Then check the receipt it leaves behind against what this page told you to expect. ``` ## Related - [Yours, provably](https://mensara.ai/how-it-works/ownership) - [Receipts](https://mensara.ai/how-it-works/receipts)