Skip to main content
The assistant is an AI chat panel beside your document in the editor. It reads the resume or letter you have open, and the job posting when the document is linked to an application, then suggests changes as edits you accept or reject one at a time. It never changes your document on its own.

Before you start

Connect an AI provider with your own API key. See Connecting an AI provider. If you open the assistant before connecting one, it offers to connect one on the spot.

Open the assistant

In the editor, select Assistant (the sparkle button) in the editor bar, just before Share, or press ⌘ J (Ctrl J on Windows and Linux). Press it again, or select Close the assistant (×) in the panel, to close it.
Resume editor with the Assistant panel open on the right, showing What should we work on? and three suggestions: Find weak bullets, Tighten to one page and Draft a summary

The assistant beside a resume, ready for a first request

On wide screens the assistant opens as a third column. On smaller laptops it takes the place of the left panel, on tablets it slides in as a drawer, and on phones it fills the screen. You can also reach it from elsewhere:
  • In the command bar (⌘ K), type a question and choose Ask the assistant. Outside the editor, this opens the document you edited last and sends the question there. See Using the command bar.
  • On an application, Prepare for next step opens its linked resume or letter with the assistant and suggestions such as How well do I fit this role?, Draft a follow-up email and Prepare me for the interview. See Managing an application.

Ask for a change

1

Pick a suggestion or write your own

A new conversation starts with a few suggestions, such as Find weak bullets, Tighten to one page and Draft a summary for a resume, or Make the opening stronger and Make it shorter for a letter. When the document is linked to an application, Tailor to the … posting appears first.Or type in the box (Ask, or describe a change…) and press Enter. Use Shift Enter for a new line.
2

Let it read and reply

The assistant reads your document, then answers in a sentence or two and shows its suggestions as a set of proposed edits. You can keep editing your document while it works.
Specific requests work best: “Rewrite my summary for a senior frontend role”, “Make the bullets in my latest job lead with results”, or “Shorten this letter to under 250 words”. When you only ask a question, such as “Which bullet is the weakest?”, the assistant answers without proposing edits.

Review proposed edits

Each proposed edit shows where it lands (for example, Summary · paragraph 1), the current text struck through, the new text highlighted, and a short reason. The same change is marked on the page, and the section shows a proposed badge, so you can read it in place.
Resume editor where the summary on the page shows the old text struck through and the new text below it, with a banner reading 1 proposed edit on this page · nothing changes until you accept, and the matching card in the Assistant panel

A proposed edit, shown on the page and in the assistant

Card titled 1 proposed edit, with the location Summary · paragraph 1, the old summary struck through, the shorter new summary highlighted in green, the reason, and Accept and Reject buttons

The proposed edit card

  • Select Accept to apply an edit, or Reject to dismiss it.
  • When a set has more than one edit waiting, Accept all applies them together.
  • With an edit focused, press A to accept, R to reject, and ↑ ↓ to move between edits.
Accepted edits show Applied and the set’s title changes to, for example, 1 of 1 applied. If you change a passage yourself before deciding, its edit shows Out of date: the text has changed since. Ask again to get a fresh suggestion. The assistant only rewrites what your document already says. When a posting asks for something your document doesn’t mention, it asks you first instead of inventing it.

Undo an accepted edit

Accepting shows a toast, Edit applied, with Undo. You can also press ⌘ Z (Ctrl Z) like any other change, or restore an earlier version from History. An edit you undo becomes pending again in the conversation, so you can accept it later. See Undoing changes and version history.

Answer a question

When the assistant needs a fact or a choice from you, it shows a blue question card with a few answers. Select one, or type your own under Or in your own words… and select Send. The assistant then continues with your answer.
A request to add accessibility work, followed by a blue question card asking whether the user has done accessibility work, with the answers Yes, I have and No, skip it and a box to answer in your own words

A clarifying question from the assistant

Choose what’s shared with each message

Above the message box, chips show what goes with your next message: the document, and the job posting when the document is linked to an application. Select the × on a chip to leave it out, including before your first message. Without the document, the assistant can’t read or edit it for that message. The line under the box always says what will be sent and to which provider, and nothing is sent until you press send. When you select Prepare from an application’s detail panel, the assistant uses that application’s posting, even when several applications share the same resume.

Attach files

Select the paperclip in the message box to attach files, such as a job description PDF or your notes. You can attach them before your first message or later in the conversation. Each file shows as a chip; select its × to remove it before sending.
  • Plain text, Markdown and JSON files are read as text (up to the first 40,000 characters).
  • Images, PDFs and MP3 or WAV audio are passed to the model directly. Whether the model can use them depends on your provider.
  • Other file types are attached but not read. Paste the relevant text instead.
  • You can attach up to 10 files per message, each up to 25 MB, and 100 MB per conversation.
Attachments work with local, S3-compatible or Vercel Blob storage and stay private. If you use a reverse proxy, allow request bodies of at least 50 MB: files travel base64-encoded, which makes them about a third larger.

Search the web

If you connect an optional Web access service in Integrations, the assistant uses that selected service for search and enhanced reading. With no external connection, supported direct OpenAI, Anthropic and Gemini configurations can supply native search through your existing AI account. See Assistant tools for model and endpoint requirements. You can also supply a public page URL: the assistant uses the selected reader or built-in reading. Page reading does not require a separate search account. Failed reading leads to pasted-text recovery; clipped or incomplete pages are identified before you rely on them. The conversation shows search/read progress, failures and retrieved Sources. Sources remain visible after reopening the conversation. A failed search keeps your conversation and does not silently switch paid providers.

Switch models

The model chip in the assistant’s header shows the model in use. Select it to switch to another connected provider for this conversation, or choose Connect another… to add one in settings.
Model menu open under the gpt-5-mini chip, listing the provider Work gateway with its model gpt-5-mini and a checkmark, and Connect another…

The model menu

Stop, retry or copy a conversation

  • Stop a reply with the stop button that replaces send while the assistant is working, or press Esc in the message box. If it stopped before proposing anything, you’ll see Stopped. No edits were proposed. with Continue.
  • When the provider returns an error, your message is kept. Select Retry, or Switch model to try another provider.
  • Copy transcript at the end of the conversation copies the messages as plain text, speaker by speaker, without the tool steps, and includes the retrieved source links.
A single reply can run for about four minutes. If it hits that limit, the assistant says “Time limit reached. Your progress is saved. Ask me to continue.”

Past conversations

Each conversation belongs to one document. Opening the assistant continues that document’s latest conversation. Select New conversation (the pencil) to start fresh, or Past conversations (the clock) to see earlier ones.
Past conversations list under This document, showing three conversations titled by their first message, each with its time and outcome such as 1 of 1 edits accepted or no edits

Past conversations for this document

Conversations are titled by your first message and show how many edits you accepted. Conversations about your other documents appear under Other documents; opening one takes you to that document. To delete a conversation, hover over it, select ×, then confirm. Its messages and attachments are deleted, but edits you accepted stay in your document. A conversation becomes read-only if its document is locked or deleted, or its provider was deleted. Unlock the document, or pick another model from the model chip, to continue.
Replies stream while the conversation is open. On instances without Redis (some self-hosted setups), a reply that’s still running when you reload the page can’t be picked up again; its finished messages are still saved.

Connecting an AI provider

Add, test and manage the providers the assistant uses.

Assistant tools

Everything the assistant can read and do.

Tailoring a resume for a job

Link a resume to an application so the assistant sees the posting.

Writing a cover letter

Use the assistant on letters, too.