Create branded PDFs and images from words, HTML, or a picture

You give thirds.ai your words, your HTML, or a picture of a design you own, and it builds an editable template in your brand. You then render that template as a PDF or as a PNG, JPEG, or WebP image, from the web app or from your own code. This article shows the three ways to start, the two kinds of output, and one short path to your first file.

Three ways to start

Every start ends in the same place, which is one editable template that you can fill with new data. The AI chat and the visual editor work on the same draft, so you can switch between them while you make changes.

Describe it with words

Open Build a template, choose your brand kit, and say what the result must show. Name the values that change between uses, such as a client name or a due date, and the AI keeps them as data fields. A new AI build costs 50 credits, and each later message in that chat costs 10 credits. Read the price beside the chat button before you send, because opening the page spends nothing.

For example, you can write "Make a one-page project brief PDF with my brand colours. Include a project name, client name, due date, summary, and a list of tasks." The chat builds the draft, renders it, and checks its own work before it says Done. Learn more on the AI template generator page.

Paste your HTML

Open Paste HTML and paste your source. The editor opens it without an AI call, so this step is free. Put a field name inside {{ }} wherever a value changes, and the Data panel fills it. If you want the AI to apply your brand, attach the HTML file in the chat, and that message uses AI credits. The HTML to PDF page shows what the renderer supports.

Import a picture you own

Open Import image, choose your brand kit, and attach one PNG, JPEG, or WebP image that you have the right to use. The AI rebuilds the layout as an editable template with your colours, logo, and fonts. The result is a rebuild, not a pixel copy, so check the text and spacing before you publish. The import costs 50 credits. Read Rebuild an image as an editable template for a full example.

Two kinds of output

One template can produce both kinds of file. You choose the output when you export or when you send an API request.

PDF

A PDF follows a page size, such as A4 or Letter, and it can run to 100 pages in one file. Use it for invoices, reports, proposals, certificates, and other documents that people print or keep. Page breaks, headers, and page numbers come from your CSS and the pdf options. See Set PDF and image options for the details.

Images

An image captures a fixed rectangle at device scale 1. Choose png, jpeg, or webp, and set a width from 320 to 7,680 pixels and a height from 200 to 4,320 pixels. The default size is 1,280 by 720 pixels. Use images for social cards, static ads, and link previews. Make Open Graph images from HTML walks through one example.

Your brand kit does the styling

A brand kit holds your colours, tone guidance, logos as PNG, JPEG, or WebP, and fonts, either an included font or an uploaded WOFF2 file. Create one in Brand kits before your first AI message, and the chat keeps it for every draft in that conversation. Read the brand kits page to see how one kit styles many templates.

This finished invoice uses the gallery template in the steps below. Copy its HTML and sample data to start with the same design.

A white invoice with blue totals and a short line-item table HTML

HTML for this image

Your first file in five steps

This path spends one credit, and a new verified account gets 25 free credits each month with no card.

Sign in and verify your account.
Open the gallery, choose Client invoice, and choose Copy to my account.
Open the Data panel and change the sample names and amounts. The live preview is free.
Press Export. A successful PDF costs one credit, and a failed one costs nothing.
Find the file in Renders, where it stays for 30 days, and keep your own copy.

If you prefer code, send your HTML straight to the API. Replace YOUR_API_KEY with your key in a private terminal.

curl --fail-with-body --silent --show-error https://thirds.ai/v1/pdf \
  -H 'Authorization: Bearer YOUR_API_KEY' \
  -H 'Content-Type: application/json' \
  -H 'Idempotency-Key: YOUR_UNIQUE_PDF_REQUEST' \
  --data '{"html":"<h1>Quarterly report</h1><p>Ready to share.</p>"}'

The response holds an id and a status. When the status is succeeded, read download.url, add https://thirds.ai in front of that path, and download the file. Create your first PDF shows the poll and download steps, and Use templates with your data shows how to send a saved template with new values.

What each action costs

Action	Credits on success
One PDF or one PNG, JPEG, or WebP image	1
Build a template in a new AI chat	50
Send a later message in that chat	10
Send a later message with a new source image	50
Edit by hand or refresh the live preview	0

Failed work costs nothing, and a render that reaches a safety limit is not billed. See pricing for plans and credit packs, and Understand credits and billing for reservations and spend caps.