What counts as a Yet
A concrete first-hand record: what you wanted, what you tried, and what actually happened. It is not a rant or a vague wish list item.
CantYet keeps track of things you wanted an AI to do but couldn't. Log them from inside your assistant, keep them private, and publish the ones worth solving — until someone, or the next model release, closes them.
Posting happens inside your AI — connect once via MCP: https://mcp-staging.cantyet.ai/mcp
What is a Yet?
A Yet is something you wanted an AI to do, but it could not do yet. CantYet keeps those moments as records, so when the capability finally arrives, the change becomes visible.
A concrete first-hand record: what you wanted, what you tried, and what actually happened. It is not a rant or a vague wish list item.
Today's failure becomes tomorrow's evidence. When a model, product, or workflow starts solving it, the timestamp shows how the frontier moved.
The assistant can record the failure while the context is still fresh: the tool, the prompt or steps, and the observed limitation.
Yets start private. Publish only the records that may help other people, vendors, or model builders understand demand.
CantYet is not just a notebook of broken prompts. It is a way to see which AI limitations disappeared, how long they stayed open, and who was waiting for them to be solved.
Four fields, enforced. Rants don't fit the template — records do.
translate all text to English, keep every layout identical. Also tried per-slide translation and pasting text boxes individually.