Most guides cover one feature at a time. This one is different: it walks through combining three separate modes — text-to-speech, image-to-video, and direct publishing — into a single workflow that turns a photo and a written script into a published video, without ever touching a traditional video editor.
Why chain modes instead of using one at a time
Each mode on its own solves a narrow problem. Chained together, they solve a bigger one: going from "I have a photo and something to say" to "it's live on TikTok" without exporting and re-importing files between separate tools.
Step 1 — Write your script and generate the voiceover
Start with text-to-speech. Write out what you want said — up to 5,000 characters — and choose a language (20 are supported), a voice (female or male), and a speed (normal or slow, useful for denser or more technical scripts that need to stay understandable). The output is a clean MP3 file, ready to use as narration.
Keep the script conversational and trim it to what actually needs saying — a voiceover that's too dense is harder to follow than one that's a little sparse, especially on a platform where people are watching with sound sometimes off.
Step 2 — Turn a photo into a video using that voiceover
Take the MP3 generated in step 1 and pair it with a single image using image-to-video. Because the video's duration locks to the audio track automatically, you don't need to manually time anything — the video will run exactly as long as your voiceover does.
A few choices that matter here:
- Enable a motion effect (Ken Burns works for most images) so the photo doesn't sit static while the voiceover plays.
- If your script has a key phrase worth emphasizing visually, add it as a word-by-word text overlay timed to roughly track the speech.
- Choose the "cover" fit mode if you're targeting a vertical format like TikTok or Reels and your source photo isn't already vertical.
Step 3 — Publish directly
Once your video is generated, publish it straight from the same screen to your connected TikTok and/or YouTube accounts — no download, no manual re-upload. Set your caption, choose the privacy level, and you're done.
When this workflow makes sense
This chain is particularly useful for:
- Announcements and quotes — a single striking image with a short spoken message, published quickly.
- Product highlights — one product photo, a short voiced description, published without filming anything.
- Recurring content formats — once you've settled on a style (same fit mode, same motion effect, same voice), repeating the workflow with a new photo and script each time is fast.
Frequently asked questions
Do I have to use all three steps every time?
No — each mode works independently. This workflow is a suggestion for when you don't have footage but do have a photo and something to say.
Can I use my own recorded voice instead of text-to-speech?
Yes — image-to-video accepts any MP3, whether it's generated by text-to-speech or recorded yourself. Text-to-speech simply removes the need to record and edit narration manually.
What if I want to combine several photos, not just one?
Generate one image-to-video output per photo (each with its own short voiceover segment), then merge the resulting clips together using the fusion mode before publishing.