AI shorts generator

Generate TikToks, Reels, and YouTube Shorts

Vertical video, generated end to end. Every model on Treza renders 9:16 natively, captions are burned in for muted feeds, and finished shorts post straight to YouTube and TikTok.

Generate a short->See pricing

9:16 on every model

Vertical is not a crop here. Veo 3.1, Sora 2 Pro, Seedance 2.0, Wan 2.7, and Grok Imagine all generate 9:16 natively, and the Sequence node exports a true 1080 by 1920 portrait frame.

Sound that ships with the frames

The current video models generate audio alongside the picture, so ambience and effects arrive in sync rather than getting layered on after. Add narration on top with a text to speech node when the format calls for it.

Captions built for the feed

Word-level timestamps from Whisper drive karaoke-style caption groups, tuned by words per caption, three styles, five vertical positions, and three sizes. The bold style exists specifically for Shorts and TikTok.

Hook, beats, payoff

A short is a structure, not one clip. The language model node drafts the beats, each beat generates as its own shot, and the Sequence node cuts them together in order with the narration mixed over the top.

Post to YouTube and TikTok

Finish the pipeline with an upload node. YouTube publishes as a Short with generated title, description, and tags, and TikTok posts through the direct post API to a connected account.

Four variations per prompt

Image and video nodes take a batch size of up to four, so one run gives you several takes of the same beat to pick from instead of one and a rerun.

How it works

From idea to finished video

Step 01
Write the hook
Give the language model your topic and format. It comes back with the beats and a shot description for each one, which is the part that decides whether the short holds.
Step 02
Generate the shots vertically
Set the video node to 9:16 and pick a model. Fast models are for iterating on structure, and the flagship models are for the version you actually post.
Step 03
Narrate and caption
Text to speech reads the script, Whisper transcribes it back with word timings, and the captions node burns in the words so the short works on mute.
Step 04
Post it, then repeat it
Publish to a connected YouTube channel or TikTok account. Add a schedule trigger and the same pipeline keeps posting on the cadence you set.

Use cases

What people build with it

YouTube Shorts at a daily cadence

Short-form rewards volume and punishes gaps. A scheduled pipeline keeps the cadence going through the weeks you do not have time to make anything.

TikTok content without a camera

Generated footage, generated narration, burned-in captions, direct posted to a connected account. No filming and no export dance.

Instagram Reels

Export the same 1080 by 1920 vertical file and upload it to Reels. Instagram is not yet a connected publishing destination, so this one is a download rather than an auto-post.

Repurposing a long video into clips

Drop the source into the timeline editor, cut the moments that matter, transcribe, and caption each one as a standalone vertical short.

Trend-reactive content

Publish the pipeline as an API endpoint and trigger it from wherever you spot the trend, so the short is rendering while the trend is still live.

Series content with a fixed format

Lock the structure in the system prompt once. Every run keeps the format and changes the subject, which is what makes a series feel like a series.

FAQ

AI shorts generator, answered

Can Treza generate true vertical video, or does it crop?

It generates vertical natively. Every video model in the catalog supports a 9:16 aspect ratio, and the Sequence node has dedicated 1080p and 720p portrait output modes. When shots of different shapes get stitched, they are scaled and padded rather than cropped, so nothing important gets cut off the edge of the frame.

How long can a short be?

A single generated clip runs up to 20 seconds depending on the model. Stitching is how you get to a full short: the ready-made templates build 30 second shorts from four beats and 60 second shorts from eight, and one Sequence node handles up to 12 shots. YouTube accepts Shorts up to three minutes, which is well within reach.

Are the captions automatic?

Yes, inside a pipeline. Wire the transcribe node into the captions node and it runs unattended on every future run. Whisper returns word-level timestamps, so captions can appear a few words at a time in sync with the narration rather than as static blocks.

Which platforms can it publish to?

YouTube and TikTok, both through official APIs with OAuth-connected accounts. YouTube uploads can be Shorts or standard videos with a visibility setting, and TikTok uses the Content Posting API for direct posts. Instagram, Reels, and X are not connected destinations, so those are manual uploads of the exported file.

Can I schedule shorts to post automatically?

Yes. A schedule trigger runs a published pipeline hourly, daily, weekly, or on a cron expression like 0 9 * * 1-5, in the timezone you choose. Each scheduled run generates a new short and publishes it, and the script node can remember previous runs so the topics do not repeat.

Do the videos come with sound?

Most of the current video models generate audio with the frames, so a rain shot arrives with rainfall and a street shot arrives with traffic. On top of that you can mix in a narration track from a text to speech node, and clip audio can be kept or muted per layer when the two would fight.

How much does it cost to make ai shorts generator with Treza?

Treza runs on prepaid credits with no subscription. Each generation is charged at the model's own rate from your balance, and only successful runs are charged, so a failed generation costs nothing. A typical video generation settles around $1.06, and credit packs start at $5 and never expire.

Can I call this as an API instead of using the canvas?

Yes. Every pipeline can be published as a versioned HTTP endpoint. Call the typed /invoke endpoint with JSON in and JSON out, or point any OpenAI SDK at the OpenAI-compatible endpoint. Swap a model on a node later and the API your product calls does not change.

Do I need to know how to edit video?

No. Start from a template, change the topic, and run it. If you do want frame-level control, the timeline editor is there with multi-track video and audio, per-clip trims, fades, and volume, but nothing about the automated path requires opening it.

Keep exploring