Designer

The Designer is an agent that turns your writing into visuals. Give it a tweet, an essay, a note or a link, and it makes:

• Essay images: one image with your words on it.
• Decks (carousels): several pages, for example a cover plus one idea per page.
• Generated images: illustrations or photo-style pictures, such as newsletter headers, X Article covers or images between sections. These have no text in them.

Every visual uses your brand kit (your colours and fonts).

The Designer does not publish or schedule anything. To post a visual, ask your X Agent, LinkedIn Agent or Threads Agent (see "How do I post a Designer carousel or image?").

You talk to the Designer in its chat. It has no queue, no weekly plan and no routines.

1. In the left sidebar, click Marketplace, then open the Agents tab. You can also click the + next to Your Agents and choose Hire an agent.
2. Find Designer (it is under Featured and the Design category).
3. Click Hire.

You land in the Designer's chat. Agents, including the Designer, are part of the All-Access plan.

The Designer greets you in its chat ("Hey {your name} — I'm your Designer.") and checks for a brand kit.

• If you already have a brand kit, it says it will use that kit, and you can send it something to design right away.
• If you don't have one yet, you see a box headed First, your brand kit. with a Build my brand kit button. Click it to start the three-step brand kit builder (see "How do I build a brand kit with the Designer?").

A brand kit is the set of colours and fonts every Designer visual uses. In the Designer's chat, click Build my brand kit (or just say "Let's build my brand kit"). The Designer shows three cards, each with a live preview:

1. Step 1 · Pick your colours: tap a palette, or tap Custom and set your own Background, Text and Accent colours. Click Use these.
2. Step 2 · Pick your font: tap a pairing (a title font and a body font). Click Use these.
3. Step 3 · Name it: type a name in the Brand kit name field and click Create my brand kit.

At any step you can type what you want instead of tapping, for example "plum" or "a classic serif", and the Designer moves on.

The new brand kit becomes your default, so new visuals start with it.

Yes. In the Designer's chat, tell it what you want and it builds the brand kit directly:

• Describe the look: "black and white with a silver accent, clean sans".
• Let it choose: "you pick" or "surprise me".
• Point to a look: "like my newsletter" with a link.
• Use your site: "use my website" plus the address. The Designer reads your site's colours and fonts and builds the kit. If it says "Could not read that site", check the address or describe the look instead.

The kit becomes your default, and the Designer tells you in one line what it chose.

Brand kits are managed in the Designer's side panel:

1. Open the Designer's chat and click its name at the top of the chat to open the side panel.
2. Open the Brand tab.
3. Under Your brands, click a brand to edit it, or click + (New brand) to add one. Under Presets, open a preset to save your own brand from it (presets themselves can't be edited).

In the brand form you can change the Heading font, Body font, Colors (Palettes or Custom), the Image style guide, and whether it is the Default brand. Click Create brand or Save brand when you are done.

To remove a brand, open it and click Delete brand. Visuals that used a deleted brand switch to your default brand.

You can also ask in chat: "use Midnight for my visuals" makes that kit your default.

You can pick any heading font and any body font from the ones brand kits offer, in any mix: they don't have to be one of the Designer's pairings. Tell the Designer ("Space Grotesk for headings, Nunito for body"), or set Heading font and Body font in the Designer's Brand tab.

• Heading font: Fraunces, Playfair Display, Inter, Space Grotesk, Lora or Geist Mono.
• Body font: IBM Plex Sans, Inter, Source Sans, Lora, Geist Mono or Nunito.

Other fonts (for example Poppins, Epilogue or a font file of your own) can't be added to a brand kit yet. The Designer offers the closest match instead. To ask for your font to be added, tell Ty (the help chat in the bottom right corner) which fonts you use, and he sends the request to the Typeshare team.

Image style is a short written guide that every image the Designer generates with that brand follows (for example the colour mood, photo or illustration, how busy it is).

• The first time you ask for a generated image and your brand has no image style yet, the Designer offers two or three looks, makes a small sample of each, and asks you to pick one. It saves your pick as the image style.
• To change it, ask the Designer ("make my images less busy", "try photos instead"), or edit the Image style box in the Designer's Brand tab.
• A one-off request ("this one at night") changes only that image, not the guide.

In the Designer's chat, send the words or say where they are. For example:

• "Turn my last tweet into a quote graphic."
• "Turn this essay into a carousel, with a cover and one main idea per card." (paste the essay or a link)
• "Make a checklist graphic from my notes on morning routines."

The Designer can use layouts such as Essay, Tweet, Apple Notes, Quote, Carousel cover, Step, Numbered list, Checklist, Pills, Photo, Full image (a picture you supply, filling the whole card with nothing on it), Big number and Call to action.

Sizes are Auto (grows with the text, single images only), Portrait (4:5, the default for decks) and Square. A visual can have up to 20 pages.

The Designer only uses your own writing: your Digital Brain (your posts, essays and notes), or text and links you share in the chat. It repeats or lightly edits your words. It does not invent hooks, stats, quotes or calls to action.

• It names its sources and shows the original passages so you can compare.
• If you paste something new, it asks before adding it to your Digital Brain.
• If it can't find anything on a topic, it says so and asks you for the text or a link.

See the Digital Brain help for how your brain is filled.

Ask in the Designer's chat. The Designer updates the same visual instead of making a new one. For example:

• "Make it warmer" or "use a serif".
• "Add my handle to every card", "hide the page numbers", "put the date top left".
• "Make card 3 a quote" or "no byline on the last card".
• "Make the text bigger".

To edit it yourself, click the visual's title or the pen icon (Open in the editor) above the pages. The editor's Theme panel lets you change the layout, aspect ratio, text size, brand, header, footer and byline.

That is how the Pills layout draws by default: every second bubble is filled and stepped in, like the other side of a conversation. While a card is drawn that way, no option under Text alignment (in Theme > Details) is picked. To line every bubble up on one side, pick Left, Center or Right there.

Each paragraph and each list item is a point and gets the next number (or a tick). Headings, quotes, and a second paragraph inside a list item don't: they keep their place with no number, in the editor and in the exported image alike. To add a plain line below the list, such as a headline, make it a heading (Text style menu or /).

Ask the Designer for a picture, for example "make a header image for my newsletter about walking" or "make a cover for my X Article".

• Shapes: wide (for X Article covers), landscape (for Substack newsletter headers and images between sections), square or portrait.
• A Substack newsletter header can go straight onto a newsletter you've scheduled: ask the Designer for a header for it.
• Generated images have no text in them.
• Each image can take up to a minute and counts toward your AI usage.
• You can ask it to follow up to 4 reference images, as long as they are already in your Media library.

For a full newsletter or X Article, the Designer can propose a header plus one image per main section. Once you agree, it makes the header first and the rest to match.

Visuals the Designer makes stay in the chat as drafts until you save them.

1. Under the image (or in the header row of a deck), click Save to Media.
2. Choose Media (the top level), one of your folders, or New folder….

The button changes to "Saved" or "Saved to {folder}". A deck saves as one item. If you changed the visual in the editor on the Write page, Save to Media saves it with your changes, even though the chat still shows the earlier pictures. If you change it again after saving, the button reads Save again (see below).

You can also say "save it" or "save it to Carousels". If you name a folder that doesn't exist, the Designer creates it. With a plain "save it", the Designer files it in an existing folder that clearly fits, or at the top of Media.

Once saved, your X Agent, LinkedIn Agent and Threads Agent can find and attach it. See the Media library help for folders.

• A deck in the chat: click the download icon (Download every page) in the deck's header row. You get a ZIP of JPEG images, one per page.
• From the editor: open the visual (Open in the editor), open the three-dot menu, choose Export, then PNG images, JPEG images or PDF document.
• From Media: after saving, use the item's menu in Media and choose Download (or Download all pages for a carousel).

Yes. Ask, for example, "Make a visual for one of my Substack posts this week." The Designer reads your content calendar (every agent's posts that are scheduled or waiting for review), picks the post you mean (or the one that would gain most from a visual), makes the visual from that post's own words, and puts it on the post.

The post keeps its time and its review status. Open it (Edit) to see the visual, swap it or remove it.

Yes, from the Long Form editor:

• Covers: in Preview & Publish, open Covers and click Generate using Designer. It draws three cover options from your piece to pick from (or click More options for three new ones). The one you pick is the Substack cover (3:2) and, when the piece goes to X, is redrawn for the X Article header (5:2).
• Graphics in the piece: type /, choose Designer graphic, describe it and pick Wide or Square. It's drawn in a box where you asked and replaces the box when it's ready.

Each picture takes about a minute and counts toward your usage (three cover options are three pictures, plus one for X's shape once you pick). If you haven't hired the Designer, the button hires it first, then makes the picture. See the writing and publishing help for the Long Form editor.

Only for newsletter headers. Your X, LinkedIn, Threads and Substack Notes agents never have the Designer make a picture on their own. When one of their posts should have media (see Media on posts on the agent's Info tab), they only use your Media library. If nothing there fits, the post goes out as text and the agent's icon appears next to the post's image button: hover it to see the picture it wanted.

To have the Designer make a visual for a post, ask for it in the post editor:

• Click the image icon and choose Make a visual. The Designer reads the whole post.
• Or select the words you want the visual to be about, then click Make a visual on the bar that appears above them.

It draws three options to pick from: brand-kit cards (for a line, a list or a number, ready in a few seconds) and illustrations (for a scene, in your brand kit's image style, about a minute each, and each counts toward your usage). Click More options for three new ones (you can say what to change first), or pick another brand kit from the menu at the bottom left and click Redraw in this kit. The one you pick goes on the post as a draft that you can remove, and it's added to your Media library only if you save it; the others are deleted.

A Substack newsletter the agent schedules without an image still gets a header from the Designer automatically. You can remove it or have the Designer draw another.

The Designer never posts. Ask your publishing agent instead:

• In a channel with the Designer and your X Agent, LinkedIn Agent or Threads Agent, ask the Designer for the visual, then @-mention the publishing agent: "@X Agent post the carousel".
• In the X Agent's, LinkedIn Agent's or Threads Agent's own chat, save the visual to Media first, then ask the agent to post it ("post the carousel I saved to Carousels").

How carousels go out:

• X: up to 4 images per post, so a longer carousel becomes a thread with one post for every 4 pages.
• LinkedIn: the whole carousel goes out as one swipeable PDF document.

Publishing agents add a Designer visual to a post only when you ask, never on their own.

• You are out of AI usage. Generated images count toward your AI usage. Open Settings from your account menu at the bottom of the sidebar, then AI Usage, to add credit or turn on auto-reload.
• Your reference images aren't in Media. The Designer can only follow reference images that are already in your Media library. Upload them to Media first.
• The website couldn't be read (when building a brand kit from a site). Check the address, or describe the look instead.
• The image model's safety filter turned the description down. The option says "The image model's safety filter turned this description down. Describe it another way and try again." In Make a visual, click More options, type a different description and click Make 3 more; in the Designer chat, ask again in other words. A refused image doesn't count toward your usage.

If you edit a visual in the editor on the Write page after saving it to Media, its button in the chat reads Save again instead of "Saved". Click Save again and the current version replaces the saved one in Media, in the same folder. The button goes back to "Saved". If the chat was already open while you edited, switch back to its tab or window and the button updates.

So you can choose. Make a visual (on a post) and Generate using Designer (on a Long Form cover) both open the same window: three different options from the Designer, each shown as it's ready. Click one and choose Use this one. For something else, click More options, optionally type what to change ("warmer colours", "no people") and click Make 3 more: you get three new ones, different from the ones you've seen. The menu at the bottom left shows the brand kit they're drawn in (your default to start): choose another of your kits or a preset, then click Redraw in this kit to redraw the same ideas in it (More options also uses it), or Make a new brand kit in the Designer chat, which opens your Designer chat to make one. Closing the window deletes the options you didn't pick. Each illustration counts toward your usage; cards don't take long and use no image generation.

Sometimes the AI that plans the Designer's options is slow to respond. When it takes too long, Typeshare stops and shows "Making options took too long. Please try again." in the window that Make a visual or Generate using Designer opens. Nothing is lost. Close the window and click Make a visual (or Generate using Designer) again. If it keeps happening, ask Ty for help so we can look into it.

Yes. In the window that Make a visual or Generate using Designer opens, the menu at the bottom left shows the brand kit the options are drawn in, starting with your default kit. Choose another of your kits or a preset: a line under the options says they use the last kit, with Redraw in this kit to redraw the same three ideas in the new one (nothing is redrawn until you click it). More options also draws in the kit you chose. To make a new kit, choose Make a new brand kit in the Designer chat: your Designer chat opens with the request ready to send. Your new kit shows in the menu next time.

The AI service Typeshare uses stopped answering for everyone for a while. It has nothing to do with your plan or your AI usage. We get an alert and fix it; close the window and click Make a visual again in a little while.