How to Add Subtitles to a Video With AI
This guide walks through each step in detail, covering browser-based tools, desktop editors, and mobile apps. The process is fundamentally the same across all tools, with differences mainly in the interface and available styling options.
Step 1: Choose Your Subtitle Tool
Your choice of tool depends on three factors: budget, content type, and where you plan to publish. For free social media captioning with animated styles, CapCut is the most practical option. For branded content with custom fonts and colors, VEED offers the deepest styling controls. For maximum transcription accuracy on professional content, HappyScribe provides both AI and human-reviewed options. For a fully free, offline workflow, Subtitle Edit combined with OpenAI's Whisper model runs everything on your own computer.
If you are unsure which tool to use, start with CapCut for short-form content or YouTube Studio for longer videos. Both are free and produce good results. You can always switch to a different tool later, and most subtitle files (SRT format) are interchangeable between tools.
Consider your video length. Browser-based tools handle files up to about 1-2 GB and 30-60 minutes of content without issues. Longer or larger files may need a desktop tool or a platform with higher upload limits. Check the tool's file size and duration limits before starting a large project.
Step 2: Upload Your Video
Most tools accept common video formats including MP4, MOV, AVI, WebM, and MKV. MP4 with H.264 encoding is the safest choice and works with every tool. If your video is in an uncommon format, convert it to MP4 first using a free converter like HandBrake.
Upload times depend on your file size and internet speed. A 10-minute 1080p video is typically 500 MB to 1.5 GB. Some tools offer cloud storage integrations (Google Drive, Dropbox) or direct URL import (paste a YouTube link) that can be faster than uploading from your local drive. If available, use these options for large files.
Before uploading, listen to a sample of your audio. If the audio quality is poor, with loud background music, heavy echo, or very quiet speech, consider cleaning it up first. Free tools like Audacity can reduce background noise and normalize volume levels. Clean audio produces significantly more accurate subtitles and reduces the time you spend on corrections.
Step 3: Generate Auto-Captions
Once your video is uploaded, look for a button labeled "Auto Subtitle," "Auto Caption," "Generate Subtitles," or similar. Most tools ask you to select the spoken language before generating. Choose the correct language, as selecting the wrong one will produce gibberish output. If your video contains multiple languages, some tools support multi-language detection, but most work best with a single primary language selected.
Processing time varies by tool and video length. Cloud-based tools typically process at 2-5x real-time speed, meaning a 10-minute video takes 2-5 minutes. Some faster tools process even quicker. Desktop tools running Whisper locally depend on your hardware, but a modern GPU can process 10 minutes of audio in about 1-2 minutes.
After processing completes, the tool displays the generated subtitles alongside your video in an editor view. Each subtitle segment shows its text content and its start and end timestamps. The segments are synchronized to the video playback, so you can watch the video and see the captions appear in real time.
Step 4: Review and Edit the Transcript
Even the best AI transcription makes mistakes. Budget 5-15 minutes of review time per 10 minutes of video content. Play the video and read along with the generated captions, pausing to correct any errors you notice. Common issues include misspelled proper nouns, confused homophones (like "their" and "there"), dropped filler words, and incorrect punctuation.
Most subtitle editors support keyboard shortcuts that make review faster. Common shortcuts include pressing Tab or Enter to advance to the next subtitle segment, pressing a key to start and stop playback, and using arrow keys to adjust segment timing. Learn the shortcuts for your tool before starting a long review session.
Pay attention to timing as well as text. A subtitle that appears a half-second too early or too late is distracting. Watch for segments where the text appears before the speaker starts talking or lingers after they have moved on to the next sentence. Most editors let you drag the edges of subtitle segments to adjust their start and end points visually.
If your content uses specialized terminology, technical jargon, or brand names, do a find-and-replace pass for common misrecognitions. For example, if the AI consistently transcribes your brand name incorrectly, a single find-and-replace fixes every instance at once.
Step 5: Style Your Captions
Subtitle styling depends on your platform and audience. For social media (TikTok, Reels, Shorts), animated captions with word-by-word highlighting are the current standard. These draw attention and help viewers follow along, especially on muted playback. CapCut and VEED both offer extensive libraries of animated caption styles.
For YouTube and longer-form content, clean, readable captions are more appropriate. White text with a semi-transparent dark background provides the best readability across different video backgrounds. A sans-serif font like Arial or Inter at a medium size works well. Position captions in the lower third of the frame, which is the conventional placement that viewers expect.
For professional and corporate content, match the caption styling to your brand guidelines. Use your brand font if the tool supports custom font uploads, and use brand colors for text or background elements. Consistency with your visual identity reinforces professionalism.
Regardless of style, prioritize readability. Ensure sufficient contrast between the text and the background, keep the font size large enough to read on mobile screens, and avoid placing captions over visually busy areas of the video. Two lines of text per subtitle segment is the maximum for comfortable reading.
Step 6: Export Your Video
You have two main export options: burned-in subtitles or a separate subtitle file. Burned-in (hardcoded) subtitles are rendered permanently into the video frame. Viewers cannot turn them off, but they are guaranteed to display correctly on every platform. This is the standard choice for social media content.
A separate subtitle file (SRT, VTT, or another format) keeps the captions as a standalone file that you upload alongside your video. The viewer's player renders the captions, and viewers can typically toggle them on or off. This is the preferred approach for YouTube (upload the SRT in YouTube Studio), web-based video players, and any platform where you want to offer optional captions.
If you need both, many tools let you export the subtitle file and the burned-in video in the same session. Export the subtitle file first (SRT is the most universal format), then export the video with captions rendered in. This gives you maximum flexibility for different distribution channels.
Check your export resolution settings before downloading. For social media, 1080p vertical (1080x1920) is standard. For YouTube, 1080p horizontal (1920x1080) or 4K (3840x2160) is typical. Some free tools limit export resolution, so verify that your tool supports the resolution you need before spending time on editing.
Troubleshooting Common Issues
Very low accuracy: If the generated subtitles are mostly wrong, check that you selected the correct language. Also check your audio quality. If the audio has heavy background music or is very quiet, try extracting and cleaning the audio track first using a free tool like Audacity.
Subtitles out of sync: If captions consistently appear too early or too late by the same amount, look for a global offset or shift feature in your editor. This adjusts all subtitle timestamps at once by a fixed amount. If individual segments are out of sync, you will need to adjust each one manually.
Missing segments: If the AI skipped sections of audio entirely, it usually means the speech in those sections was too quiet, too fast, or obscured by other sounds. You can add subtitle segments manually in most editors by clicking at the point in the timeline where the gap occurs and typing the missing text.
Export quality issues: If the exported video looks blurry or the captions are hard to read, check your export resolution settings. Also verify that the caption font size is large enough. Text that looks fine on a desktop preview can be illegible on a phone screen if the font is too small.
The AI does the heavy lifting of transcription and timing, but the review step is where you ensure quality. Budget a few minutes of review time per 10 minutes of video, focus on proper nouns and technical terms, and always preview the final export on a mobile screen before publishing.