State-of-the-art AI Font Generation Model

Official Mixfont Python SDK

Examples | Supported platforms | How it works | Model inputs | Usage

The official Python SDK for the Mixfont AI font generation API. It lets you create AI-generated font files from Python applications and scripts.

Mixfont is a frontier AI lab developing generative AI for fonts. The Mixfont font generation model creates complete, web-safe TTF font files from a natural-language prompt or a public reference image, so applications can turn generated lettering, sketches, logos, or visual references into editable type instead of a flat image. Fonts generated via the API are unique and licensed for commercial use.

For more information, see the Mixfont website and the full Mixfont documentation.

Examples

These examples show prompt and image inputs paired with generated font files from Mixfont.

Input	Generated font preview
generate a font for a soccer team logo. Make the letterforms soccer themed, bold, and unique.	Download Cipher-Striker-Ultra.ttf
	Download Wuthering-Heights-Display.ttf
	Download Gossamer-Editorial-Serif.ttf
	Download Dystopian-Brush-Stroke-Display.ttf
	Download Dissonant-Wave-Sans.ttf
	Download Zephyr-Ink-Script.ttf

Supported platforms

• Python >= 3.9
• Standard CPython environments with network access

How font generation works

Font generation is asynchronous. Start a generation with exactly one input:

• prompt: a text description of the font to generate.
• image_url: a public HTTPS reference image for the style you want the model to follow.

The create call returns a generation id and, when available, a polling URL. Use mixfont.generations.wait(...) for built-in polling, or call mixfont.generations.get(...) yourself until the status reaches succeeded, failed, or cancelled. When a job succeeds, ttf_url contains the generated TTF download URL.

Model inputs and outputs

Use text generation when you can describe the type direction, such as category, style, use case, spacing, contrast, or distinctive details. Use image generation when a visual reference is the clearest source of truth, such as a sketch, sign, logo, poster, screenshot, or existing design mockup.

Reference images should be publicly reachable HTTPS URLs that point to JPEG, PNG, or WebP files up to 20 MB. Clear images with readable letterforms, strong contrast, clean edges, and cropped text regions generally produce better results.

Generated font files are returned as TTFs. Download or persist the returned ttf_url after the job succeeds, then rehost the file in your own storage before using it in production. Returned TTF URLs are temporary and will be deleted within 24 hours.

Create an API key

To start making calls to the API you'll need an API key. If you don't already have a Mixfont account, visit Mixfont and sign in.

Go to the Developer Console to create your first API key. When calling the API, include the key in the x-api-key header for each request.

Make sure to copy the API key to a safe place. The API key will only be visible in the dashboard once. If you lose your key, you'll have to create a new one.

Installation

Install it from PyPI:

pip install mixfont

Usage

For the full API documentation, see Mixfont docs.

Import the package:

from mixfont import Mixfont

Instantiate the client:

import os

mixfont = Mixfont(api_key=os.environ["MIXFONT_API_KEY"])

Create a font generation:

generation = mixfont.generations.create(
    prompt="A condensed sci-fi display font",
    glyph_set="standard",
)

print(generation["id"])

Fetch the generation later:

generation = mixfont.generations.get("generation_id")

print(generation["status"], generation["progress_percent"])

Or wait for the generation to finish:

result = mixfont.generations.wait(generation["id"])

print(result["ttf_url"])

Create a generation from a reference image:

generation = mixfont.generations.create(
    image_url="https://example.com/reference.png",
)

API

Constructor

mixfont = Mixfont(api_key, timeout=30.0)

mixfont.generations.create(...)

Starts a new font generation and returns immediately.

Provide exactly one of prompt or image_url.

Glyph sets

standard includes English letters, numbers, and basic punctuation. extended supports all Latin languages, including special characters, and costs more API credits.

mixfont.generations.get(generation_id)

Fetches the current status of a generation.

mixfont.generations.wait(generation_id, ...)

Checks the generation until it reaches a terminal status.

wait returns the completed generation when it succeeds. It raises an error if the generation fails, is cancelled, or times out.

Best practices

Write specific prompts that describe the type category, visual style, intended use case, and distinctive details.

Start with standard when comparing directions, then use extended once you have a candidate worth testing more deeply.

Store the generation id, original prompt or image URL, and glyph_set with each result so your team can compare outputs later.

Test generated fonts in real content, including headings, numbers, punctuation, labels, and the longest strings your product needs to support.

Keep your API key on the server and read it from an environment variable such as MIXFONT_API_KEY.

Development

PYTHONPATH=src python3 -m unittest discover -s tests
python3 -m build

Links

• Website
• GitHub
• X