Frontify Brand Assistant
Frontify is a platform for brand guidelines, DAM (digital asset management), and templates. Its Brand Assistant is a chat in the guidelines. A user can ask it to review copy, for example against the tone of voice, and it answers from the text of the guidelines.
Frontify controls the look of work in a different way. Its templates lock the logo, the colors, and the typography, and an admin unlocks the parts that users can change. Frontify also has a Brand Check app that analyzes assets against the guidelines, and Frontify calls that app an experiment.
| Group | Automatic checks |
|---|---|
| Where the work must be | A chat in Frontify guidelinesA user uses the Brand Assistant inside a guideline in Frontify. The user must be logged in and must have access to that guideline. Only an account admin can enable or disable the Brand Assistant for the account. An admin must also activate it for each guideline. The Brand Check app runs on assets in a Frontify library or project. It runs when a user creates a new asset or changes the workflow status of an asset.1 |
| Can an agent run it? | No: its MCP server only reads guidelinesThe Frontify MCP (Model Context Protocol) server has tools that read guidelines, brand colors, fonts, and assets. Its list of all available tools has no tool that checks work against the guidelines. Frontify describes brand compliance checks as a use of its MCP (Model Context Protocol) server. The outside AI tool retrieves the guidelines and then reviews the work itself.6 |
| File types | Text onlyThe Brand Assistant works with text. It uses only the text-based content of the guidelines as its source. Digital templates and print templates export PDF, PNG, and JPEG files. Motion templates export MP4, GIF, and HTML5 files.1 |
| What the check gives back | Answer with referencesA response from the Brand Assistant includes references to the guideline content that it used. The Brand Check app gives a report for each asset, with visualizations, issues, and ratings.1 |
| Who decides | Advises in chatThe Brand Assistant gives answers and suggestions in a chat. Frontify tells users to verify each response with the references. An admin can set required approvals on a template. Then a user must request approval before the user can export a publication from that template. If a user edits a publication after approval, the user needs approval again before the user can download it.1 |
| Where the rules come from | Guideline text in FrontifyThe Brand Assistant takes its answers from the content of the brand guidelines in Frontify. It does not rely on pre-trained knowledge. Frontify tells customers to write key rules, for example tone of voice, compliance, and messaging, as text in the guideline. The assistant does not infer or guess. Editors and Owners can change the name, the welcome message, and the disclaimer of the Brand Assistant for each guideline.1 |
| Brands in one account | Several brands in one accountOne Frontify account can hold several brands. The MCP (Model Context Protocol) server has a tool that lists all brands in the account.6 |
| Plan with the check | By quoteFrontify states that the Brand Assistant is generally available and enabled by default for customers. It is no longer in beta. A customer can use templates only if templates are part of the Frontify plan of that customer. Asset audit trail is an enterprise feature.1 |
| Free plan | Not published |
What does Frontify check?
| Check | Result | What the vendor says |
|---|---|---|
| Logo | Locked in templates | Frontify states that its templates lock in key brand elements such as logos. Teams can change content only inside the limits that the template sets.2The vendor's words
|
| Color | Locked in templates | In a template, an element keeps its color unless an admin unlocks the color. The admin can then restrict which colors the user can choose.3The vendor's words
|
| Type | Locked in templates | In a template, text styles stay locked unless an admin unlocks them. If the admin unlocks the typeface, the user can choose from the typefaces in the source file.3The vendor's words
|
| Copy and terms | Reviews copy in chat | A user can ask the Brand Assistant to review copy against the guidelines, for example against the tone of voice.1The vendor's words
|
| Claims and legal lines | Not published | The vendor's pages do not say. |
The fine print
Limits that Frontify states
Text-based only: Images, PDFs, and videos are not currently used.
The Brand Assistant uses only text-based guideline content. It does not use images, PDFs, or videos.1
If important information only exists in these formats, it won’t be used.
The Brand Assistant does not use PDFs, images, videos, page titles alone, or Marketplace guideline blocks. If a rule exists only in one of these formats, the assistant does not use it.9
Content that exists only in translated versions is not used.
The Brand Assistant answers in more than 100 languages, but it always uses the default language version of the guidelines. It does not use content that exists only in a translated version.1
In public portals, the Brand assistant is only visible to logged-in users.
In a public portal, only logged-in users can see the Brand Assistant.1
We don’t currently support approvals for Video Publications in Frontify.
Frontify does not support approvals for video publications.8
An asset can only have one active approval request at a time
An asset can have only one active approval request at a time. An approval preset for templates and publications is the exception.14
Frontify's Content API doesn't yet support semantic search or navigation hierarchy, and access to block-level data is limited.
Guideline retrieval through the MCP (Model Context Protocol) server is limited. Frontify states that the agent can fail to find guidance that exists.7
Multi-page PDFs aren't read reliably.
Through the MCP (Model Context Protocol) server, an agent cannot read multi-page PDFs reliably. Frontify tells customers to put such content in the guideline as text or images.7
Grounding reduces error, it doesn't eliminate it.
The MCP (Model Context Protocol) server returns brand data, and the outside AI tool interprets it. Frontify states that the AI tool can misread or invent.7
This led us to our first experiment, the Brand Check app
Frontify calls the Brand Check app its first experiment in automatic review of assets.17
Specs is still in development.
Frontify describes Specs, an AI agent that reviews assets in a project and leaves comments. Frontify states that Specs is still in development.17
Which plans does Frontify have?
Frontify prices by MAU (monthly active users) and publishes no fixed prices. Frontify proposes a package for each customer.12 Frontify no longer creates new self-service plans. A customer who needs a plan change must speak with the sales team about the enterprise offerings.13
Prices are from the vendor's pricing page.
What does an example look like?
This example is from the pages of Frontify.
- In an example from Frontify, a user asks the Brand Assistant: “Review this text based on our tone of voice.”1
What else do Frontify's pages say?
- All template layers start locked
- When an admin uploads a template to a template library, Frontify locks all layers. The admin then unlocks the elements that users can edit.3
- Approval requests for assets
- In a project, a user can request approval of an asset from one or more reviewers. The result can change the workflow status of the asset automatically.14
- Audit trail with CSV export
- The asset audit trail records events for an asset, for example approval requests, revisions, downloads, and workflow statuses. A user can export it as a CSV file.11
- MCP server at no extra cost
- The Frontify MCP (Model Context Protocol) server is off by default. An account admin can activate it at no additional cost. The default tool pack is read-only.15
- A second assistant organizes assets
- Frontify also has a Brand Assistant in its DAM (digital asset management) libraries. It manages and organizes assets, and it is off by default. It works with 100 selected assets at most.16
- 100+ languages
- The Brand Assistant supports more than 100 languages and answers in the language of the prompt.1
- No training on customer data
- Frontify states that it does not use customer data to train AI models. Azure can process and keep interaction data for up to 90 days for monitoring and quality purposes.1
Who does it suit?
It suits a team that already keeps its guidelines and assets in Frontify. People ask the Brand Assistant about the rules while they write. Teams that make the same layouts many times can use the locked templates, with required approvals before export.
If your code must fetch the guidelines or the files, The Brand API Register describes the Frontify API.
When is another product a better fit?
- If you need an automatic check of finished images or documents, look at Bynder or Rainbrand. The Brand Assistant uses only text, and Frontify calls its Brand Check app an experiment.
- If an agent must run the check, note that the Frontify MCP (Model Context Protocol) server only reads. The agent retrieves the guidelines and then reviews the work itself.
- If your main need is long text in many writing tools, look at Markup AI or Writer.
Compare
Where are the vendor's own pages?
Sources
- help.frontify.com: brand assistant in guidelines checked November 16, 2026
- frontify.com: templates checked November 16, 2026
- help.frontify.com: setting up templates checked November 16, 2026
- help.frontify.com: understanding print digital and motion templates checked November 16, 2026
- help.frontify.com: how to enable or disable the brand assistant checked November 16, 2026
- mcp.frontify-integrations.com: admin checked November 16, 2026
- help.frontify.com: how to use the frontify mcp use cases and prompts checked November 16, 2026
- help.frontify.com: create approvals approve publications checked November 16, 2026
- help.frontify.com: get the best results from the brand assistant checked November 16, 2026
- help.frontify.com: getting started with frontify templates checked November 16, 2026
- help.frontify.com: asset audit trail checked November 16, 2026
- frontify.com: pricing checked November 16, 2026
- help.frontify.com: manage billing subscription self service checked November 16, 2026
- help.frontify.com: feedback and approval process checked November 16, 2026
- help.frontify.com: frontify mcp checked November 16, 2026
- help.frontify.com: brand assistant in dam checked November 16, 2026
- frontify.com: brand check with specs the reviewer that knows your guidelines checked November 16, 2026