Video & motion studio
Make short videos, slideshows, and reels from a few photos or a simple idea — captions and voiceover included, no editing experience needed.
Who it's for
- Anyone who wants video content but has never used a video editor.
- Small teams and solo creators who need to post often and don't have time to edit.
- People who already have photos, a voice note, or a podcast and want to turn them into something shareable.
What you can do
- Type a topic and get a finished vertical reel with narration and on-screen captions.
- Turn a set of photos into a smooth slideshow with music or a voiceover.
- Upload a video and get captions added automatically, ready for social.
- Turn an audio clip into an eye-catching "audiogram" with a moving sound wave and your brand colors.
- Animate a single still image into a short moving clip.
- Drop in a podcast or long recording and pull out the best moments as ready-to-edit posts.
Every finished video is saved to your workspace Library, so you can drop it straight into a post.
Getting started
Open the video tools
Head to the AI section and pick the tool that matches what you have — a topic, photos, an audio clip, or a video.
Add what you've got
Type your idea, upload your photos, or attach your audio. The studio handles the script, narration, and captions for you.
Generate and review
Let it build your video, then preview it. Everything lands in your Library when it's done.
Post it
Attach the finished video to a post in the Composer and publish.
TIP
New to AI features? Start with Video generation for a gentler walkthrough of turning ideas into clips.
For developers
Most tools render the video deterministically with ffmpeg — there is no native text-to-video model in the pipeline. Scripts, hero images, narration, and transcripts come from the Qevron AI gateway (text, image, TTS, STT); ffmpeg does the actual encode. The only exception is the Talking Brand Avatar, which calls an external image-to-video provider and is therefore ENV-gated.
Every tool requires an authenticated user, an active subscription, and the relevant feature flag. Output videos are saved to the workspace Library as standard /uploads assets you can attach to posts.
Tools and endpoints
| Tool | Endpoint | Input | Output | Feature | Credits |
|---|---|---|---|---|---|
| AI Reel | POST /ai/video/reel | topic (required), aspect, language, voice | 9:16 narrated, captioned .mp4 | ai.video | 20 |
| Audiogram Studio | POST /ai/video/audiogram | audio file or text, voice, aspect | branded waveform .mp4 | ai.video | 10 |
| Motion Composer | POST /ai/video/slideshow | slideFileIds[] (required), durationSec, transitionSec, aspect, audioFileId/voiceoverText, captionText | xfade slideshow .mp4 | ai.video | 10 |
| Auto-Captions | POST /ai/video/captions | video file (required), language, burn | .vtt + caption-burned .mp4 | ai.video | 10 |
| Talking Brand Avatar | POST /ai/video/avatar | imageUrl/imageFileId/heroPrompt, motion, provider, aspect, durationSec | image-to-video clip | ai.video | 30 |
| Podcast-to-Posts | POST /ai/podcast-to-posts | audio file (required), count, language | carousel design drafts | ai.template | 10 |
All paths are under the API prefix (/api/v1). File uploads are capped at 60 MB. Credits are charged after a successful render.
AI Reel — Give it a topic. The studio writes a short spoken script, generates a vertical hero image, synthesizes narration via TTS, then burns synced captions over the image and encodes a 9:16 .mp4. Brand voice and accent color are pulled from the workspace brand kit when available. No video model is used — the reel is composed with ffmpeg.
Audiogram Studio — Upload an audio file, or pass text to synthesize narration via TTS. The studio renders an animated waveform .mp4 using your brand background and accent colors. Uploaded audio is transcribed (best-effort) to add captions; default aspect is 1:1.
Motion Composer — The /motion-composer page lets you order existing images by file id into a timeline. ffmpeg renders an xfade slideshow between slides, with optional voiceover (voiceoverText via TTS, or an uploaded audioFileId) and burned captions. Tune durationSec and transitionSec per slide.
Auto-Captions — Upload a video; the studio extracts the audio, runs STT, and builds caption cues — timed when the gateway returns segments, otherwise evenly distributed. You get a .vtt sidecar returned inline plus, unless burn=false, a caption-burned .mp4.
Talking Brand Avatar — Animates a still image into a short clip via an external image-to-video provider. Source the still from imageUrl, imageFileId, or a generated brand hero (heroPrompt). Duration is clamped to 3–8s (default 5), aspect defaults to 9:16.
| Provider | Value | Default model | Required ENV |
|---|---|---|---|
| fal.ai | fal (default) | fal-ai/veo3/fast | FAL_API_KEY |
| Runway | runway | gen3a_turbo | RUNWAY_API_KEY |
If the selected provider's key is missing, the endpoint returns 503 "not configured". Set AI_VIDEO_PROVIDER (default fal) to choose the default.
Podcast-to-Posts — Upload long audio (podcast, webinar, voice note). The studio transcribes it via STT, extracts up to 6 share-worthy highlights (count, default 3), and generates one editable carousel design draft per highlight in your workspace. Output is a set of draft designs, not a video.