Artificial Intelligence · Audio/Video Development · Image Recognition

Build Your Own AI Image & Video Studio on a Free API

By 雨夜寻晴天 · Jun 25, 2026

Read original on juejin.cn ↗ Google Translate ↗ Alt translation

Why it matters

A free, indefinite API for text, image, and video generation is rare — most free tiers are limited trials. This project shows how a developer can turn that into a self-hosted, privacy-respecting alternative to expensive subscriptions, and it signals a growing trend of API providers betting on developer ecosystems rather than direct consumer sales.

Summary

Agnes AI made its text, image, and video model APIs free indefinitely — no token limits, no time bombs. In the first week, the image model alone generated over 2 million images. That caught the attention of a developer who was tired of paying for Midjourney, Runway, and Pika for occasional use.

The result is Agnes Creator Studio, an open-source Gradio 6.0 web interface that bundles text-to-image, image-to-image, text-to-video, image-to-video, and multi-image keyframe animation into a single local tool. It supports common aspect ratios (including TikTok's 9:16), resolutions up to 1080p, and adjustable frame rates. All generated content is saved with a history feature.

The project is MIT-licensed and runs on Python 3.10+. Deployment options include local setup via pip, Docker, or one-click deployment to Hugging Face Spaces. The developer notes that while the tool won't compete with polished commercial products, its value is being free, usable, and fully under the user's control — with the caveat that Agnes's free API policy could change in the future.

Key takeaways

— Agnes AI offers free, indefinite API tokens for text, image, and video generation — no time limit or usage cap announced.

— The image model generated over 2 million images in its first free week.

— Agnes Creator Studio is an open-source Gradio 6.0 web interface that wraps these APIs.

— Supported features: text-to-image, image-to-image, text-to-video, image-to-video, and multi-image keyframe animation.

— Image-to-video requires a public URL for the source image unless deployed on a server.

— Output supports resolutions up to 1080p, frame rates from 12 to 60 FPS, and aspect ratios including 16:9, 9:16, 1:1, 4:3, and 3:4.

— The project is MIT-licensed and can be deployed locally, via Docker, or on Hugging Face Spaces.

— The developer acknowledges that Agnes's free API policy may change in the future.

Our take

The project highlights a third path for AI tools beyond expensive subscriptions and restrictive free tiers: developers building their own interfaces on top of free APIs.

The choice of Gradio over a more polished framework signals that rapid prototyping and simplicity are prioritized over production-grade UX — a pragmatic trade-off for a personal tool.

The developer's candid admission that the API might not stay free forever adds a layer of realism that many open-source AI projects gloss over.

This pattern — free API + open-source wrapper — could become a common model for niche AI use cases that don't justify a full SaaS subscription.

Concepts & terms

Gradio

An open-source Python library for quickly building web-based UIs for machine learning models and APIs, popular for prototyping AI demos.

Keyframe Animation (in AI video)

A technique where multiple images are used as keyframes, and the AI generates smooth transitional video frames between them, useful for product showcases or creative storytelling.

Image-to-Image

An AI process that takes an input image and transforms it according to a text prompt (e.g., changing style or content), with adjustable strength to control how much the output deviates from the original.

Source: juejin.cn ↗ Google Translate ↗ Backup ↗