LittleGuy keeps one memory of what matters to you: what you told which assistant, and when. Every AI you use starts from the same ground truth.
Works with any AI that speaks MCP.
A memory example
A saved memory is useful when you can find it, inspect it, and change it. Here is the path LittleGuy is built for.
Verified with a test account
The fictional release meeting is Thursday.
One correction, history kept
The fictional release meeting is Wednesday.Corrected to Thursday.
This example was saved and recalled across two MCP test clients on the live service. Its sources and correction history were verified in the account export.
Tell Claude a preference or decision and ask it to save that memory in LittleGuy.
Ask a second connected MCP client what you saved. Its answer can point back to the saved item.
Open the memory in LittleGuy to see what was saved and when.
Replace a stale memory with the current version, or forget one you no longer want used.
Review connected clients and remove access when you no longer use one.
LittleGuy does not train models on your memories. You can inspect, correct, export, or delete what it stores.
Every memory belongs to your account. Another account cannot read or change it.
Inspect saved memories, their sources and the record of changes.
Disconnect an OAuth client or revoke an access key from your account.
Download your account as JSON, or delete it after confirming with your email.
Sign in, connect an agent, and follow the guided steps.