5.7.0 — Release Notes
Changelog for Agnes Video Generator: new features, improvements and bug fixes for every release.
What's New
Features & Improvements
- Multi-API-Key rotation & throughput boost — configure multiple Agnes API keys (env + config, auto-dedup); requests rotate across keys via a
KeyRing, and a 429 retry-on-key-switch mechanism instantly switches to the next key on rate-limit instead of backing off. Layered rate limiting gives video submissions a dedicated1×Key/minbucket alongside the shared20×Key×0.8/minbucket, roughly doubling achievable throughput with 2 keys. - User-supplied scene reference images (creative) — in creative long-video mode you can now upload a reference image per scene; the pipeline uses your image instead of AI-generated storyboard frames for that scene.
- Task deletion (UI + API) — delete completed task folders from the task list or via
DELETE /api/tasks/{id}; running tasks are protected with a clear error. The whole.working_dirstays clean. - API-key list & per-key removal — the settings panel now lists all configured keys with source badges (env/config); config-sourced keys can be removed individually without clearing everything.
- Windows one-click start —
start.batfor native Windows usage without Docker.
Refactoring & Optimizations
- Frontend engineering re-architecture — the single-file
index.htmlwas refactored into a Vue 3 + Vite + TypeScript project (frontend/) with build output committed tostatic/. This brings type safety, modular components/composables/i18n, and a stable dev/build workflow (verified by CI's frontend-build job). - Unified image normalization module — new
utils/image_normalizer.pynormalizes reference/end-frame images to target size with black-bar padding before model submission, improving i2i/keyframes input consistency.
Bug Fixes
- API-key plaintext leak —
GET /api/config/keysnow returns masked keys (sk-xxx...xxxx) with stable IDs instead of plaintext, preventing key disclosure via the API; deletion uses the stable ID. - Legacy task API compatibility — the legacy
POST /api/tasksroute now forwardsscene_reference_images, fixing CI failures for older clients. - JSON robustness — optional
json_repairfallback for LLM JSON responses with trailing commas / missing colons. - npm packaging — exclude runtime-generated
static/generated/from the published npm package.
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