Choosing what the AI uses
Once your sources are ingested, you can control how much of them Edithly draws on when answering questions or generating a visual.
Use from document
In the chat composer, the document context controls (such as Use from document) let you point the AI at a specific attached source instead of your whole chat history. This keeps answers and generations focused on the material that actually matters for your question.
Generate from different starting points
When you open a visual type in the Visuals studio, you typically choose what it should be generated from:
- From the current conversation — uses the chat (and anything discussed in it) as context
- Based on a topic — describe a topic directly; no document or chat history required
- From an attachment — uses a specific document you've attached to the chat
Picking the right starting point keeps generations relevant — for example, basing a Cheatcode on a specific PDF rather than the whole conversation if that PDF is what you actually want summarized.
@ mentions
Type @ in the chat composer to open a mention picker and reference:
| Mention type | Use for |
|---|---|
| Visual | Point at a generated visual and ask Edithly to refine or extend it |
| Document | Point the AI at a specific attached document |
| Member | In group chat, notify or address a teammate |
Selecting a visual from @ can open a preview; while it's open, your next message can automatically include that visual as context — useful for iterating ("make this mindmap simpler", "add a slide about X to this presentation").