Studio & tools

Every tool is a step-by-step wizard. The user fills it in, presses generate, and a background worker renders the video. Output lands in the Library, ready to download or publish.

How generation works

1. The user completes a tool wizard and presses Generate.
2. A job is queued and the page shows a live progress bar (%).
3. The background worker renders with FFmpeg + the AI providers.
4. On success the asset appears in the Library and a “your video is ready” email is sent.

Tools

Idea → Video

The flagship tool. The user types an idea (or imports a Reddit question), Tokify writes a script, splits it into scenes, picks stock footage, adds an AI voiceover and dynamic subtitles, and stitches it with transitions. Steps: Format → Details → Subtitles → Script → Voice → Generate. Voice is chosen from the ElevenLabs picker (filter by language, accent and gender, with in-language preview).

Avatar editor

Turns a script into a talking AI avatar video via D-ID or HeyGen. The user picks a presenter and a voice; the provider generates the speech and lip-sync together. The admin chooses which provider the editor uses (Auto D-ID→HeyGen, D-ID, or HeyGen). Five edit styles (Impact, Cinematic, Paper, Pulse, Animated) apply a complete look (grade, zoom, vignette, captions). Can also caption an uploaded clip.

Animate (photo → video)

Upload a single photo and AI turns it into a short, moving clip — make the subject dance, twerk, jump, sing, wave or catwalk, or describe the motion yourself. Powered by fal.ai image-to-video with two selectable models: PixVerse v5 (fast & cheap) or Kling 2.6 Pro (higher quality). Steps: Photo → Effect → Generate. Duration (5/8s for PixVerse, 5/10s for Kling) and resolution are picked in the wizard. Unlike the render tools this is billed as a fixed, admin-configured minute cost per clip (not by output seconds), because each fal render costs real money. Needs the fal.ai API key configured in Settings → API keys. JPG, PNG and WEBP are accepted.

Clips

Cut a long video into multiple short, vertical clips. Configurable clip count and target length, optional burned-in subtitles, and a preview-before-save review grid: trim each clip, then keep the ones you want.

Subtitles

Upload a video; Tokify transcribes it (Whisper) and burns styled, animated subtitles. Choose any saved caption style.

Translate

Upload a video, pick a target language, and get a subtitled translation burned in with your chosen style.

Narrator (text-to-speech)

Paste text, pick a voice, get an audio file. Numbers and dates are spelled out before TTS so the voice never stumbles.

Reddit tool

Same pipeline as Idea→Video, but seeded from a pasted Reddit link — the question and body are imported automatically.

Caption styles

Users build and save their own caption styles (font, size, colour, position, outline, highlight/karaoke, entrance animation) in the futuristic style builder, then reuse them across tools. Word-level karaoke highlighting is synced to Whisper timestamps; phrases are grouped on natural punctuation so lines read well.

Library

Every generated asset (video, audio) is stored per user with a title, dates, duration and a poster thumbnail. Users can search, sort, preview, download, publish/schedule (see Connections) or delete. Deleting removes the file and all its dependencies from the server.