7.2.0 — Release Notes
Changelog for Agnes Video Generator: new features, improvements and bug fixes for every release.
What's New
Features & Improvements
- External asset mode with Jimeng (Dreamina) collaboration. A third execution mode next to auto and manual. At each checkpoint the app renders a ready-to-copy creation prompt for the selected tool — picture, audio and duration constraints included — so the storyboard stays consistent while the shots come from Jimeng's video model with native audio.
- The videos checkpoint prompt is the storyboard script itself, numbered per scene with its target duration, so it can be pasted into Jimeng scene by scene.
- The videos checkpoint validates that every scene clip is present before you continue, and reviews the imported stills at the references checkpoint.
- The tool list comes from a registry (
GET /api/providers); Jimeng is the only entry for now and custom tools are intentionally not user-definable. - Batch import with normalization and probing. Select several mp4/png files at once: clips are scaled and padded to the task resolution, re-timed to 30 fps with H.264/AAC, and stills are normalized to PNG. Each file is reported back with duration, resolution, frame rate, whether it carries an audio track, and its video/audio codec. File names map to scenes by the first number they contain (
scene_1.mp4→ scene 1), falling back to upload order, with an explicit mapping available. - Subtitles can follow narration or on-screen dialogue. In external asset mode the subtitle text source is selectable: the narration script, or per-scene dialogue that the screenwriter generates alongside the video prompts — useful when the native audio is character speech.
- Keep the original audio track. When the video's own audio is used, the app skips TTS and silent-audio placeholders entirely, keeps each clip's audio in the final cut, and derives the subtitle timeline from the real per-clip durations.
- The collaboration mode is labeled beta, with the official docs one click away. The external-asset entry points carry a Beta bubble, and both the create panel and the checkpoint collaboration card link the official workflow guide and FAQ (including how to handle Agnes
503/video queue is full), in all 22 interface languages. - Storyboard prompts follow the input language. In external asset mode the scene prompts are written in the same language as the user's idea instead of being forced into English, so the prompts handed to Jimeng match what was asked for.
Bug Fixes
- Subtitle burning no longer silently drops subtitles when the system ffmpeg lacks libass. The system binary is probed for the
subtitlesfilter and, when missing (for example Homebrew's ffmpeg formula), the bundled static ffmpeg is used instead of degrading to a plain, subtitle-free concat. - Imported file names can no longer land outside their scenes. A name carrying a year or timestamp (e.g.
20261015.mp4) is no longer mistaken for a scene index and written to a non-existent scene directory. - Clips without an audio track are accepted on import and no longer highlighted as an error: whether the clip carries audio is the user's call, and the final cut falls back to the project voiceover when it does not.
- Chinese input stays Chinese for story and storyboard in external asset mode, where the language pinning previously produced an English story from a Chinese idea.
- End-frame scene indices are bounded, closing a crafted request that could exhaust worker memory before a task was even created.
- Compatibility: the new execution mode is additive — auto and manual behaviour is unchanged, and tasks created by earlier versions keep running with their saved mode.
About Agnes Video Generator
A completely free, open-source AI video generator: turn text into multi-scene AI videos with narration and subtitles.
- Free and open source — no payment, no high-end GPU required
- Text-to-video, image-to-video, digital anchor and image generation
- Powered by free Agnes AI models, rendered entirely in the cloud
- Runs via Web UI, Docker or npm, and supports self-hosting
Related Pages
Release Notes
Changelog for Agnes Video Generator: new features, improvements and bug fixes for every release.
Try Online Demo | Agnes Video
Enter a prompt and generate AI videos and images instantly
Text to Video API — Free AI Video Generation API, No GPU | Agnes Video
Programmatically access Agnes AI's free video models. Text-to-video, image-to-video, and narration APIs. Developer-friendly, zero-cost, no high-end GPU required.
Why you can trust this
Written by a practicing developer and cross-checked against authoritative primary sources such as arXiv papers, vendor technical reports, and the Stanford AI Index.
SandGrid@lcy362
Author of Agnes Video Generator · Full-stack Developer
Independent developer and author of Agnes Video Generator, an open-source (MIT) AI video generation tool built on Agnes AI's free video models. Has helped 1,000+ creators produce AI videos at zero cost. Focused on making video generation models accessible and production-ready; content is based on hands-on practice and primary research.
View on GitHubLast updated：2026-08-21
Ready to Start Creating?
"Making world-class AI belong to everyone." — Bruce Yang. It's completely free, no credit card, and you won't need a high-end GPU. Your first AI video starts at zero cost. Want to use Agnes AI's free video models? This is the easiest way in.
Clone the GitHub repo and launch in 2 minutes