Mesh

Mesh is the AI agent inside your Workspace, working in the same room, on the same things, as you. Chat is where you talk to it, and what it makes is saved in the Workspace, not buried in the conversation.

Start from what you already have

Point Mesh at something you already saved and it knows exactly what you mean — no uploading it again, no pasting it back in. What does this chapter say about limits? with the PDF pointed at. Quiz me on this with a Note. What is wrong here? with a snip of the Board. Then say where the result should end up: saved as a Note, filed under a Topic, drawn on the Board.

You never have to point. Mesh can find anything in the Workspace on its own. Pointing just shows it exactly what you mean, so the answer comes back faster and surer — Chat shows the move.

One request per kind of content

There is no fixed list of requests: say what you need. Here is one example for each kind of thing in the room, and where the result lands.

  • What do my sources say about the deadline? — Mesh searches your Library, reads the passages it finds, and names the source in its answer; a read like this saves nothing.
  • Save what we just worked out as a Note. — Mesh writes the page and it lands in Notes as an ordinary Note, editable from then on like one you wrote yourself.
  • Rename Week 1 to Foundations and move it under Term 1. — Mesh reshapes the Map itself, and that Topic's Hub page wears the new name too.
  • Put each open question on its own sticky note on the Board. — the notes land on the Board page you have open, as movable and editable as the ones you place yourself.
  • Look at the whiteboard photo I saved and tell me what it says. — Mesh opens that Image and answers from the picture, not from the name and comment it started with.
  • Summarize this lecture video. — Mesh says plainly that it can't watch or transcribe a Video, and asks for the part that matters as a transcript, a quote, or a picture instead.

The pages ahead

See what Mesh is doing

While Mesh works, you can watch a short play-by-play: searching, reading, creating, adding to Topics. Those steps show progress, not proof: when something actually changes, a compact outcome card appears after the change really saved. History & outcomes owns the full trust loop.