Repository navigation
Replies: 1 comment
|
模型设计时就是不支持视频输入的,关网页端APP端什么事 |
0 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
背景 / Current Situation
目前 DeepSeek 网页版(chat.deepseek.com)和 APP 支持上传图片和文档,但不支持直接上传视频文件。想让 DeepSeek "看"一个视频,只能手动截图逐帧上传,或者先用第三方工具转录音轨,非常麻烦。
期望 / Proposed Enhancement
希望 DeepSeek 后续能支持视频阅读(视频理解):
使用场景 / Why this matters
类似的视频理解能力在其他主流助手产品中已逐渐普及,希望 DeepSeek 也能考虑支持,谢谢!
English summary: We hope DeepSeek (web/app and agent forms) will support video reading — allowing users to upload video files (mp4/mov) so the model can understand visuals, timeline, and speech, and answer questions about the content (summaries, key points, Q&A at specific timestamps). Currently only images and documents are supported, which forces manual frame-capturing or third-party transcription.
All reactions