[FEATURE] Support importing local SRT/VTT subtitle files

[ngxccc]

Before you begin

• I have searched existing issues to avoid duplicates
• I am willing to contribute a PR for this feature

Is your feature request related to a problem? Please describe.

For videos without native CC/subtitles, users currently rely solely on the "AI Subtitles" server API. When users generate their own local subtitle files (e.g., via local Whisper or Google Colab ASR), there is currently no way to import those .srt or .vtt files directly into Read Frog's video player to translate them.

Describe the solution you'd like

Add an "Import Local Subtitles (.srt / .vtt)" feature in the Video Subtitles Settings Panel:

1. Allow users to select a local .srt or .vtt file from their computer.
2. Parse the local subtitle timestamps and text fragments into subtitle state.
3. Utilize Read Frog's configured translation providers (such as Ollama / Local LLM / Microsoft Translate) to translate and render bilingual subtitles seamlessly on the video.

Which area(s) would this feature affect? (Select all that apply)

Browser Extension

Additional context

This feature would enable users to bypass server quota limits and leverage custom ASR models (e.g., Faster-Whisper on Colab or Local GPU) while retaining Read Frog's excellent bilingual subtitle rendering and translation pipeline.

[frogGuaGuaGuaGua]

This is a valid use case. I checked the current subtitle implementation and confirmed that local imports are not supported today:

• subtitle sources are currently limited to native and ai;
• the settings panel exposes AI subtitle generation and subtitle downloads, but no file input;
• the existing SRT utility is export-only—it serializes subtitle fragments but does not parse SRT/VTT files.

The downstream pieces are reusable: the scheduler already renders time-coded { text, start, end } fragments against the active video clock, and those fragments can go through the existing translation and bilingual-rendering pipeline.

A safe first version would:

1. accept UTF-8 .srt and .vtt files in the subtitle settings panel;
2. parse and validate them entirely in the browser without uploading the file;
3. bind the imported track to the active video for the current tab/session and clear it on navigation;
4. treat it as a third subtitle source so configured providers, bilingual rendering, and translated-subtitle export continue to work;
5. show a clear error for malformed or unsupported cues.

One scope question before implementation: should the first version support only YouTube, where Read Frog currently has dedicated subtitle controls, or every supported HTML5-video platform? That choice affects where the import action is exposed and how the imported track is tied to video/navigation lifecycle.

[ngxccc]

Thank you for the quick and detailed feedback!

Regarding the scope question: prioritizing YouTube for the first version (v1) makes total sense. YouTube is the primary platform where Read Frog's subtitle controls and user base are most mature. Focusing on YouTube first will keep the scope clean, allow fast implementation, and make testing straightforward.

Expanding to other HTML5 video platforms can then be handled seamlessly in a follow-up v2 release.

The proposed v1 workflow (parsing locally in the browser, treating imported tracks as a third local source, and binding to the active video session) sounds excellent! Thanks again!

[mengxi-ream]

Hi, we have supported ai subtitle and it's in beta. You can apply here https://tally.so/r/7Rzb96

Then you can get subtitle even for the videos without built in captions

[ngxccc]

Hi @mengxi-ream, thank you for sharing the beta access link!

I have actually already joined and experienced the AI Subtitles beta (it works great!). However, the quota is capped at 400 minutes maximum. Once those 400 minutes are exhausted, there's currently no alternative way to generate AI subtitles for videos without captions through the extension.

That was precisely what led me to propose this feature! Allowing users to import their own local .srt / .vtt files (which can be generated for free via local Whisper or Google Colab ASR) gives heavy users a way to continue translating videos after hitting the quota limit, while also saving server resources for Read Frog.