Add and use Workspace documents
Bring a document into a Workspace so In Parallel — and any AI assistant you've connected — can work from it. In Parallel isn't a file library: a document enters a Workspace when you use it to propose goals, and from then on it's attached to that Workspace.
Before you start
You need access to a Workspace. If you haven't created one yet, see Create a Workspace.
Have a
.pdf,.md, or.txtfile ready — for example a project brief, a strategy doc, or a plan.
Add a document to a Workspace
You add a document through the goal-proposal flow. In Parallel reads it to suggest goals, and keeps the document attached to the Workspace.
1. Open the Add Goal modal
In your Workspace's Goals area, click Add Goal.
2. Choose "suggest goals from a document"
Under Or let AI suggest goals, click From text or documents.
3. Upload your file, then propose goals
Click Upload .pdf / .md / .txt and select your document (or paste text directly into the box). Click Propose goals. In Parallel reads the document and drafts candidate goals from it.
4. Accept or dismiss the proposed goals
Review each candidate goal and Accept the ones you want (setting an owner and deadline) or Dismiss the rest. Accepting a candidate turns it into a real goal in the Workspace. See Add a goal for the full goal-creation walkthrough.
The document stays attached to the Workspace even after you've handled the proposed goals.
Use a document with a connected AI assistant
Once a document is attached to a Workspace, a connected assistant (Claude, ChatGPT, Cursor, or Microsoft 365 Copilot) can work with it over MCP:
List them — ask "What documents are in this Workspace?" The assistant uses
list_workspace_documentsto return each document's metadata (name, type, upload date).Re-extract goals — ask it to propose goals again from a document you uploaded earlier; it re-runs extraction without a re-upload (
re_propose_goals_from_document).Upload and extract in one step — a write-enabled assistant can upload a document and propose goals directly (
upload_document_and_propose_goals), subject to the per-tool approval you've set.
Note: list_workspace_documents returns metadata only — names, types, and dates — not the full file contents.
What's next
Add a goal — the full goal-creation flow, manual and AI-assisted.
Where your execution data lives — how documents fit alongside goals, decisions, and Findings.
In Parallel MCP tools reference — the document tools and everything else an assistant can use.
What the AI can and can't see — the access boundaries around a connected assistant.