Make a YouTube video
- Write a title and a first 30 seconds that earn the click and the stay
- Produce a thumbnail that does its job
- Know which CF tool runs each step, and where YouTube itself still limits you
Step 1 — Title and thumbnail concept (Spark)
explainOn YouTube the title and thumbnail are the product before the video is. Spark generates title variants and the thumbnail concept (the three words, the one image idea). Decide these FIRST — they shape the video, not the other way around.
Run Spark for 10 titles. Pick one. Write the 3-word thumbnail line.
Step 2 — The hook and script (Cut)
explainThe first 30 seconds decide everything. Cut's Faceless builder drafts the hook and the full narration script from your outline — spoken-word rhythm, not essay prose. You read it back and cut anything that doesn't sound like you talking.
Have Cut draft your hook and script. Read the first 30 seconds out loud.
Step 3 — Thumbnail art (Design)
explainDesign's Graphic mode builds the thumbnail to the concept you set in step 1 — bold type and shapes, your three words, or a photo you upload yourself. There is no AI photo generation live in CF yet, so a purely photographic thumbnail is still bring-your-own. Shrink your draft to phone size — does it still read?
Build a thumbnail in Design. Shrink it to phone size — does it still read?
Step 4 — Record and edit (Cut)
explainCut records your camera or screen and stitches it, or its Faceless builder renders the whole thing from the script — stock footage, narration, burned-in captions, server-rendered MP4 — no separate editor needed. The Faceless path is capped by your plan (30-180 seconds depending on tier), which fits a Short cleanly; for a longer long-form upload, record and edit with Cut's plain camera/screen path, which has no such cap.
Record or Faceless-render your video in Cut, matching the format to your length.
Step 5 — Description and tags (Spark)
explainSpark helps with the searchable opening line and title/thumbnail concept from step 1. There isn't a dedicated CF tool for the full YouTube description and tag list yet — write it yourself from the angle Spark already gave you. Don't keyword-stuff — say what the video actually delivers.
Write a 3-line description using Spark's opening line as your anchor.
Step 6 — Publish and cross-post (Cut's upload + Pilot + Push)
explainCut keeps every finished render in your "Your videos" list — open, re-download, or delete anytime; that is where the master file actually lives, not Idea Shield, which only locks short text notes. Cut can connect to YouTube and upload directly, but that path is founder-gated today: uploads through it land Private-only until Google finishes reviewing CF's app, and it isn't open to creators yet — for now, download the finished file from Cut and upload it yourself through YouTube Studio. For the cross-posts (the community post, the X/LinkedIn note), Pilot plans the cadence across platforms; Push actually auto-posts what it can reach today — Telegram, Discord, LinkedIn, and Bluesky.
Download your finished video from Cut, upload it to YouTube yourself, then let Push schedule your cross-posts.
Front-load the click
recapTitle and thumbnail (Spark, then Design) before the script (Cut), record and render in Cut, cross-post with Push, Alfred holding the thread. The one real limit: CF's own YouTube upload isn't open to creators yet — you still upload there yourself.
- Decide title + thumbnail (Spark, then Design) before you script.
- Cut writes the hook/script and records or Faceless-renders the video, captions burned in.
- CF's own YouTube upload is founder-gated today — you still publish to YouTube yourself.
- Pilot plans cross-posts; Push actually auto-posts what it can reach; Alfred keeps the promise.