Compare
Three ways to give an assistant a memory.
The important differences are simple: where the memory lives, whether configured AI agents can follow a shared source, whether you can edit it yourself, and what happens if you switch tools.
| Characteristic | Memory built into each app | A hosted memory service | Query me |
|---|---|---|---|
| Where it is stored | Inside that company’s product | On the provider’s servers | A folder in storage you already own |
| Shared across different AI tools | No, each app keeps its own | Usually, through integrations | Configured supported agents can follow the same user-owned documents |
| Can you read and edit it yourself | Partly, through a settings screen | Through the provider’s interface | Yes, they are ordinary documents |
| If you stop using the product | Memory remains inside that app | Depends on the provider’s export and access terms | The files stay in your own storage |
| Ongoing cost | Part of that subscription | Another monthly subscription | Free bootstrap or $19 one-time Starter; no Query me subscription |
| Who maintains it as tools change | The vendor, for that app only | The provider | The user maintains the free bootstrap; Starter releases document what changed and was retested |
Query me is not the first product to address shared AI memory, and does not claim to be. The next page separates what has been measured from what still needs testing.
See what has actually been tested.
Measured results, limits, and current compatibility are kept together so the claims are easy to check.