A brand context API returns what a model needs to write or design as a specific company. That includes who the company is, how it sounds, what it sells, what it looks like, what to avoid, and which files to use. You call the API before you call the model. Then you put the result into the prompt, or you give it to an agent.
"Brand Context API" is also the name of one product: GET /v2/context/{domain} from Brandfetch. It returns a profile of any company from the website of the company. This site covers that product and seven other endpoints that return parts of the same information. Each endpoint was called for the same company, linear.app.
What do you send?
The input depends on whose brand it is.
- Any company, by domain. Brandfetch, Context.dev, and Logo.dev take a domain. Brandfetch and Context.dev also take an email or a page address. Firecrawl takes the address of one page. You do not set up anything first, because these APIs read the public website.
- A brand that you or your customer added. Rainbrand takes the ID of a brand that someone added from sources such as a website, a brand guide, or files. A Brief also takes the task, in plain words. A Brand Skill returns the whole brand.
What comes back?
| Part | Example from linear.app | Fields page |
|---|---|---|
| Name and description | "The product development system for teams and agents." | Name and description |
| Voice | "precise", "measured", and avoid "breathless hype" | Voice and tone |
| Positioning | Three audience segments, five products, four competitors | Positioning |
| Look | Colors, fonts, type sizes, button styles | Colors, Type |
| Files | Logos for light and dark backgrounds, banners, font files | Logos, Images and files |
| Rules | Write "Linear" with a capital L. Keep the logo artwork intact | Rules and claims |
| Date or version | Read on 16 October. Brand Skill version e16b6239… |
Dates and versions |
No endpoint returned all of these parts. The coverage chart shows which endpoint returned which part.
Is that not a logo API?
A logo API returns files, for example to show the logo of a company in a list. The API of a digital asset manager returns the stored files of a company, with their rights. A brand context API returns what a model needs to act, and that is mostly words: descriptions, voice, and rules. Some products do more than one of these jobs. The Brand API Register compares brand APIs product by product, with plans and limits. This site shows what their brand context endpoints send back.
Where does the Brandfetch Brand Context API fit?
It is the only observed endpoint on this site that returned voice, positioning, products with prices, and competitors in one call. Observed means that the API reads the public website. The Brand Context API returns no color values and no files, but the Brandfetch Brand API does. The page Brandfetch Brand Context API alternatives shows what else covers the same information.
How do you pass it to a model?
Pick the fields that the task needs. Write them as instructions, and keep the date. The recipes show how, in TypeScript and Python. The templates give a schema and a prompt to copy.
Last updated 12 November 2026