Log in
WorkflowBy HypernaturalVersion 1, updated

Green screen news spot

hn/workflow-green-screen-news-spot

Put a person's own webcam read in a TV news studio: a sourced story and script, the read recorded and cut to its best takes, the person cut out with a free local matte (no green screen needed), a generated news-set backdrop with a city window and a video wall, the presenter scaled and lit into the set, and a broadcast chyron (kicker, station bug, name and headline straps, LIVE bug, a ticker already scrolling) with an optional skyline opener, in 16:9 and 9:16. About a dollar of stills; the cut, matte and graphics are local. For a narrated explainer with no presenter, use workflow-news-article-to-vertical-short.

SKILL.md

The person brings a story (or a beat), a webcam and themselves; or a read they already recorded. You deliver a news spot of them as the anchor in a TV studio: the set behind them, a broadcast chyron with their name and the headline, a ticker, in 16:9 and 9:16. No green screen is needed; the matte is made from the footage. About a dollar of stills; everything else is local. Record the story, the picks and the spend where you keep the job's notes.

Steps

  1. The story (llm-source-a-news-story): picked with the person, source/article.md written. Skip it for a comic spot about their own news (a launch, the office's team); the facts still come from what they tell you.
  2. The script (guide-make-a-script-from-source): an anchor's read opening on the hook (llm-write-a-hook), then the story's facts, the anchor's sign-off; the name strap's and headline's words written with it. Gate: the script and the strap copy.
  3. The read: the person records it on their webcam, several takes are fine; face lit from the front, a still camera, a plain wall behind if they have one. Then the best takes (guide-cut-best-takes, route A). Gate: the rough cut.
  4. Cut them out (python-track-face-and-hands): the person cutout, free. Look at a sheet of it over magenta and over the set (its Check it): a light wall leaves a pale rim round the hair on a dark set, and where hair or a gap by the neck lets the room through, repair the matte there, or pay for birefnet-matte on that span.
  5. The set (guide-design-a-set, nano-banana-design-empty-set): an empty news studio at the delivery's size, a desk edge low in frame, a video wall or a window with the story's city behind, room on one side for the strap. A real city or venue follows guide-depict-real-people-and-places. Optional: an opener, the title over the city still with a slow push and a short sting. Gate: review_choose of two sets.
  6. Compose (hn4a-compositor): the set, then the presenter cutout scaled up and raised until they sit at the desk the way an anchor does (head about a third from the top, shoulders cut by the desk), then the chyron (hn4a-compositor-chyron): kicker, bug, the headline strap, the name strap as they introduce themselves, LIVE and a place for any "on the scene" part, the ticker full from its first frame. Captions, if wanted, above the chyron (guide-caption-a-video). Gate: the live spec (hn4a-compositor, Live spec, then render) with stills at each strap.
  7. Deliver 16:9 and, from the same build with --vertical, 9:16; loudness to −16 LUFS, −14 for social (ffmpeg-normalize-loudness); per the core skill.

Choices this workflow makes

  • A matte from the footage, not a key. Webcam reads aren't shot on green; the free matte held on a close-up, and a key would eat skin and hair.
  • The presenter bigger than the webcam framed them. At the webcam's own scale they sat small and low in the set; scaled 1.18× and raised 100 px (on 1280×720) they read as an anchor.
  • The ticker scrolls from the first frame, already full of items: a crawl that started empty and filled in read as a mistake.
  • The jokes live in the straps and the crawl for a comic spot ("ANCHOR, REPORTER, ENTIRE STAFF"); the read stays straight.

Dead ends

  • Broad colour removal to clean the matte: it ate into skin. A narrow seam of the wall by the neck was fixed with a local smooth repair of the matte's edge on that span only.
  • A paid video matte that wasn't available in the connection: the free local matte did the job; check the model exists before planning on it.
  • A first title that claimed more than had happened (a series "victory" before it was won): titles say only what the source says.
  • The first 0.8 s of the read showed the person settling in their chair: the studio reveal starts after it, the opener's music carrying into the first words.

Worked example

A comic sports-desk segment from two webcam recordings (119 s and 70 s, 1280×720 at 30 fps, their sound starting 0.26 and 0.21 s in), the person as both the studio anchor and the reporter on the scene.

PartPictureOver it
Openerthe city's skyline at dusk, generated, slow pushthe headline and "SPECIAL REPORT", a sting
Studiothe anchor at the desk, cut out, 1.18× and raisedkicker, station bug, the headline strap, then the name strap; the ticker
On the scenethe same person against a stadium exteriorthe name strap again with a joke sub line, LIVE and the place
Back to the studiothe anchor's wrap"BACK TO THE STUDIO", the ticker running out

The rough cut was 10 pieces from both files, the pauses kept where the jokes needed them. Notes on the way: the ticker had to start full; the presenter larger and centred at the desk; a seam of wall by the neck repaired. Stills about $0.06–0.09 each; the matte, cut and graphics ran locally.

Related workflows

FAQs

Put a person's own webcam read in a TV news studio: a sourced story and script, the read recorded and cut to its best takes, the person cut out with a free local matte (no green screen needed), a generated news-set backdrop with a city window and a video wall, the presenter scaled and lit into the set, and a broadcast chyron (kicker, station bug, name and headline straps, LIVE bug, a ticker already scrolling) with an optional skyline opener, in 16:9 and 9:16. About a dollar of stills; the cut, matte and graphics are local. For a narrated explainer with no presenter, use workflow-news-article-to-vertical-short.

Install Hypernatural Desktop, a free Mac app, and connect it to Claude Code, Codex or Claude Desktop. Then describe the video you want and your agent finds the workflow itself, or paste this page's link (https://hypernatural.ai/skill/hn/workflow-green-screen-news-spot) and it installs hn/workflow-green-screen-news-spot and follows it.

It runs Nano Banana Pro. Hypernatural Desktop has every one of them built in, so there are no API keys to set up, though you can bring your own.

The workflow is free. You only pay for the AI generations it runs, and your agent shows you the price of every batch before it spends anything. Editing, captions and titles run on your Mac at no cost.

It calls 17 smaller skills for its steps, including Llm source a news story, Make a script from source and Cut best takes. Each one is listed on this page, and your agent loads them as the job needs them.

Yes. Ask your agent to adjust it for your job, or to save its own version. You can publish your version under your own handle so anyone can install it.

Go make something cool.

Install the app, open your agent, describe the video. That's the whole onboarding.

Sticker HypernaturalSticker BannerSticker TVBoy