A simple 4-step workflow: choose the video → automatically create clips → translate each clip → download one final SRT.
VIDEO → SRT
1 · Choose videoUpload or paste a direct link
2 · Create clipsAutomatic timestamps
3 · TranslateWatch, pause, type
4 · Export SRTOne final subtitle file
1
START HERE — Choose the full video
You only need to do this once. Upload the complete video below, or paste a direct video link if the file is too large to send.
☁️ Shared project workspace
For two-device collaboration, the deployed site needs a real cloud database/authentication service. This version keeps the project structure ready for shared login, but a plain HTML file cannot synchronize browser data between different devices by itself.
Use the same project code on both devices once a cloud backend is connected. Your imported video is not copied into local browser storage.
☁️ Cloud account & shared project
Connect this app to your Supabase backend. Use only the Project URL and Publishable key here. Never put a secret/service-role key into the website.
Not connected
Connect first, then create an account or sign in.
No cloud project loaded.
1. Gemini API key
For this private/local tool, the key is kept only in this page's memory. Do not hard-code a key into a website you give to the client. For production, put the key on a server.
Video source
Choose the video source
Use Gallery for normal videos, or Browse Files for MOV/MP4 files in Downloads, Files, WhatsApp, SD card, etc.
No video selected.
Browse device files uses the browser's native file picker when supported and starts in Downloads. If the browser does not support it, Android may use its normal picker.
Best: a direct downloadable video URL. A normal webpage/SwissTransfer/Google Drive page may require a login or may block browser access. If the link allows browser access, the site downloads it into memory and then treats it exactly like an uploaded video.
No video link loaded.
2
AUTOMATIC CLIPS — Let the site split the video
After the video loads, the site analyzes the speech in order, detects the language of each spoken section, and creates timestamp-aligned work segments. Italian is translated automatically; Himba and Herero are left for the human translator.
What happens next
The full video is divided into short clips.
Every clip keeps its exact original timestamp.
You work through the clips one by one below.
🎬 Translator workspaceeasy mode
Upload the video and run the language analysis. The segments will then appear in exact chronological order.
3
TRANSLATE — Watch each clip and type the English meaning
Italian segments are translated automatically. Himba and Herero segments stay empty for the translator. Everything remains in the original chronological order with the detected speech timestamps.
4
FINISH — Create the final SRT
When you are ready, use the button below. The site first proofreads every English translation, removes speaker labels, separates clearly labeled speakers, validates the timestamps, and then creates one clean professional SRT. It will not get stuck on the old “finish one translation” check.
Complete your translations first. The final step will proofread the English, remove speaker labels, separate clearly labeled speakers, validate timestamps, and then create the SRT.
Advanced / AI tools
These tools are optional. The simple translator workflow above is the recommended way to have a fluent Himba/OtjiHimba speaker verify the meaning.
Gemini API key
Add as many short clips as you want. Added clips stay in this list until you delete them. Set the exact original-video start/end time for each clip, then generate a single ZIP pack containing the clips and a ready-to-use translator page. Your translator only opens the page, presses play, types the English meaning under each clip, and downloads the final SRT.
No clips added yet.
Your translator does not need to import the videos. Build the Translator Pack after you finish editing the list; the ZIP contains the videos and translator page together.
Advanced: AI video analysis
Waiting for a video and API key.
Advanced: voice-note reviewer
Your Himba/OtjiHimba speaker can watch the video and record the English meaning in their own voice. The voice note records the microphone only, not the video's audio track. Use headphones so the original speech does not leak from the phone speaker into the microphone.
Reviewer recorder00:00
Choose a video above, then press Start voice note while the reviewer explains the meaning.
Normally leave the video unmuted and use headphones. Muting the video is optional; the microphone recording is separate from the video's audio.
⚡ Quick Translate Mode (legacy)advanced
Upload the full video once. The website automatically divides it into sequential time segments. Your Himba speaker plays one segment, then records or types the English meaning for that exact segment. The original full-video timestamps are automatic — nobody has to screen-record or type timestamps.
Create a clip queue after the video loads.
5. Review subtitles
The table below is the important step: your fluent Himba/OtjiHimba speaker should correct the original-language transcript and the English meaning here. The downloaded SRT uses the Verified English column, not the raw AI draft.
Final VERIFIED SRT
Important: Himba/OtjiHimba is a low-resource language. AI output can be substantially wrong. Do not deliver the draft directly. Have a fluent Himba/OtjiHimba speaker verify the source transcription and English meaning, then export the VERIFIED SRT.
Why this version is different
MP4/MOV inputdirect video analysisHimba transcriptEnglish translationtimestampsSRT export
Gemini's current video API accepts MP4, MOV and other video formats and can analyze the audio and video stream. Its File API supports much larger files, so the same workflow can be used for the client's long recording rather than manually converting every video first.