Chat with Files, Notes, Tasks & Goals in Sapience
If its already in Sapience, you can use any Agent to chat with it or work with it.
Bringing Your Files, Notes, Tasks & Goals Into a Chat
Last Updated: 2026.07.07
Overview
You've already got work living in Sapience — uploaded files, notes, tasks, and goals. When you're chatting with an agent, you often want it to look at one of those things: "summarise this note", "draft an update based on this task", "turn these three files into a briefing".
Now you can hand any of them straight to the agent. Attach something already in Sapience to your conversation, and the agent reads its full content — no re-uploading, no copy-pasting, and it works with every agent.
Think of it as saying "here, look at this" and pointing at anything already in your workspace.
Upload New vs. Choose Existing
When you click the attach (📎) button in the chat box, you'll see two tabs:
TabUse it when…What happensUpload NewThe content is on your computer and not yet in SapienceYour file is uploaded, saved to your workspace, and its content goes into the chatChoose Item Already in SapienceThe thing you want is already in SapienceYou pick it from a list; its content goes into the chat — nothing is re-uploaded
This article is about that second tab — pulling in what you already have.
What You Can Attach
From the "Choose Item Already in Sapience" tab you can pick any of these:
TypeWhat it isFileAny document you've uploaded — PDF, Word, spreadsheet, text, and moreNoteA Sapience note (page)TaskA task from your projectsGoalA goal from your projects
The picker searches across all your projects, and it's multi-select — grab a file, two notes, and a task in one go if you like.
How to Do It
- In the chat box, click the attach (📎) button.
- Switch to the "Choose Item Already in Sapience" tab.
- Search and select the items you want — from any project, as many as you need.
- Confirm. Your selections appear as removable chips above the message box (tap the ✕ on a chip to drop one before sending).
- Type your message and send. The agent now has the full content of everything you attached.
What Happens Behind the Scenes
When you send that message, here's what Sapience does before the agent replies:
- It reads each item you attached — the actual file content, or the full note/task/goal.
- It checks your access to each one. You can only ever attach something you're allowed to see; there's no way to pull in a resource you don't have access to.
- It places that content directly into the conversation as context the agent has already read.
- The agent answers using it immediately — it doesn't need to go searching for the item, because the content is already right there in front of it.
Because the content lands in the conversation itself, it stays available for the rest of that conversation — the agent keeps "seeing" your attached items on every follow-up message, not just the one you attached them on.
Good to Know
- Works with every agent. There's no special mode or setting to turn on — any agent you chat with can use attached items.
- Attach more any time. You can add extra items on a later message; they're added to what the agent already has, not swapped out.
- The agent won't re-fetch it. Since the content is already in the conversation, the agent treats it as read and gets straight to your request.
- A bad pick won't break your message. If one item can't be loaded for any reason, the agent simply carries on with the rest — your turn still goes through.
Attaching vs. Sharing — What's the Difference?
Attaching puts an item's content into one conversation so an agent can work with it right now. It doesn't change who can see the item, and it doesn't move or copy anything — it's a "look at this for this chat" gesture.
If instead you want to give another person ongoing access to a file, note, task, or goal, that's sharing — a separate feature.
Summary
The "Choose Item Already in Sapience" tab lets you drop any of your existing files, notes, tasks, or goals — from any project, several at once — straight into a chat. The agent reads them in full, respects your access permissions, and keeps them in view for the whole conversation. It's the fastest way to say "work with this" without uploading anything twice.