How ImTalking Works
Four steps, all running locally on your PC — nothing you make here ever leaves your machine. Here's exactly how to go from a still photo to a finished talking-avatar video.
Pick or import a source photo
Your library shows every photo you've imported. Click one to select it, or bring in a new photo with ＋ Import to Library.
- Use a square, well-lit, front-facing photo with a clean background — this is the single biggest factor in output quality.
- The importer lets you drag to reposition, scroll or pinch to zoom, and drag the corner handles to fine-tune the crop before saving.
- Once saved, the photo appears in your library instantly and is ready to select.
Choose how the avatar will speak
ImTalking supports two input modes — pick whichever fits what you're making:
- Text to Speech — type what you want said and ImTalking generates the voice for you. Best when you don't have a recording, or want to iterate quickly on wording.
- Audio File — bring your own MP3 or WAV. Best when you already have a voiceover, narration, or a specific recorded performance you want lip-synced.
Using Text to Speech
Type your script, search the voice picker for a voice you like, and click Generate Speech Audio to preview it before rendering the video.
- Keep individual sentences under 15 words — ImTalking splits your text by sentence and renders each one separately, so shorter sentences render more predictably.
- Use the Pitch and Rate sliders for subtle tone changes — small adjustments (±10–20) sound natural; large ones can sound artificial.
- Star your favourite voices so they surface at the top of the picker next time.
Using an Audio File
Click Browse for MP3 / WAV…, choose your file, and preview it with the built-in player before generating. There's nothing else to configure — ImTalking handles the rest.
Fine-tune Processing Options
Sensible defaults are already selected, so this step is optional — but a few settings are worth knowing about:
- Enhance – using GFPGAN applies face restoration to every frame for a sharper result, at the cost of extra processing time.
- Still mode keeps the head stable with minimal random movement — good for presentation-style videos.
- Expression Scale controls how animated the face is; 1.0 is natural, 1.2–1.5 is more expressive.
Generate your video
Click Generate Video. You'll see frame-by-frame progress as ImTalking renders each sentence and assembles the final clip.
- Click Generate Video again while a job is running to queue another — it starts automatically once the current one finishes.
- The queue keeps going even if you close the window — you'll get a Windows notification when each video is ready.
- Every finished video is saved locally and shows up in your Projects tab for quick access later.