Feature 06 // Create content

Video Toolkit

Paste a link. Get back a scene-cut, AI-voiced, captioned clip — with hook variants, viral copy, and hashtags — formatted for TikTok, Reels, or YouTube. Or point it at any web page and it records a scrolling video with voiceover. No editing app. No template. Just the result.

Try it — Get Started Read the full guide

Use cases // What people ask for

Real requests, run end to end.

Make 5 TikToks from this 10-minute YouTube video

Turn my podcast episode into 3 Instagram Reels

Extract clips about productivity from this 1-hour talk

Convert this landing page into a scrolling demo video

Find the best YouTube video about compounding interest and make a TikTok from it

Turn this recording into clips, captions, hooks, and a thumbnail idea

What you'll see

Chat-based requests — no separate UI needed

Progress indicator showing which step is running

Output folder link once done

Preview card with the final MP4

Video Toolkit // Two Pipelines

Repurpose a video, or turn a page into one.

Flow A — Repurpose any video into clips

Download (source video) → Scene detection (natural cuts) → Transcription (audio to text)
            → AI dialogue (voiceover script) → Voice render (local
              TTS) → Format & assemble (aspect ratio) → MP4

Flow B — Turn any web page into a video

Open page (real browser) → Record scroll (frame capture)
            → Assemble footage (timeline) → Voice render (local
              TTS) → Format (aspect ratio) → MP4

9:16 vertical — TikTok / Reels / Shorts16:9 — YouTube1:1 square — Instagram

Output // Every run produces this

Not a clip. A full content pack.

From one video link →

Upload-ready MP4 — scene-cut to length

Burned-in captions styled to the video

Thumbnail extracted at the best frame

AI-written voiceover, rendered with local voice

3 hook variants — curiosity, stat, direct challenge

Viral title, 2-sentence caption with emoji, 8 hashtags

Subtitle file for manual editing

Voices — 10 Local, No API Key

Atlas · Orion · Caden · Reed · Finn — five male voices

Nova · Luna · Sage · Iris · Ember — five female voices

All rendered on-device — no external speech service

Say which voice you want, or Celestos picks one

Image Generation // On-Device, In Chat

Generate images inside chat. No subscription.

Type /image followed by a prompt in Chat or Orb. Celestos generates the image on your machine — no cloud service, no account, no per-image cost. The result appears inline in the conversation.

Image rendered inline in the chat bubble

Download and fullscreen controls below every image

Run variations on the same prompt

Use a reference image as the base (img2img)

Try this now

> /image a neon cityscape at dusk, cinematic
              lighting

> /image portrait of an
              explorer, warm studio light, realistic

Get Started Free

Tips // Getting the most out of it

Field notes from real use.

Default is 9:16 vertical, ~60s — say the platform or ratio you want

Smaller output = faster processing

No URL? Just describe the topic and Celestos finds the best-scoring video automatically

Say which voice you want, or leave it out — a default is picked for you

Try this now

> Make a 60-second vertical TikTok from this
              YouTube link

> Record a scrolling video
              of my landing page — 30 seconds, horizontal

Get Started Free