Versely
    Guides

    CLI tools: terminal caption presets, then a hook pack

    Look first. Openings second.

    Versely Team••4 min read

    Look first. Openings second. A CLI product on a muted feed is recognised by type, not by a cinematic terminal that never existed. Lock the caption look — black card, monospace, position — before you buy a pack of hooks. Otherwise every opening arrives in TikTok chrome and you spend the week restyling it.

    The product is a look. The test is a set of first lines. Two purchases.

    The look is a caption preset, not a prompt

    Do not prompt a model to "make it look like a terminal." That sentence is a vibe. The durable look is a preset: font, fill, card, outline, position. Caption styles is 45 Caption Studio presets — 9 families, 5 variants each — previewed as the actual style fields.

    The family that already behaves like a terminal is Ink: white text on a black rounded card. Variants: square corners (ink_square), a 70% translucent card (ink_sheer), navy (ink_navy), top of frame (ink_top). Classic is the everyday social stroke. Hype is not a compiler. Pick Ink, or Classic plus a font. Do not invent a tenth family in a prompt.

    Position is top, center, or bottom. Terminal chrome in a 9:16 frame is a preset choice, not a thing Veo will typeset.

    Courier is in the font registry

    list_caption_fonts returns the registry by category: sans-serif, serif, display, handwriting, monospace. Courier New is in that last group — id courier, preview "Code Style". Pass font_id. Do not hope a generate invents a mono.

    Add a text overlay burns a headline you wrote: one fixed line, whole clip, not a transcript. Call the font list first. Optional background block. That is the CLI badge and the "works in your terminal" claim. Auto-captions transcribe speech. If nobody said the flag name out loud, transcription cannot print it.

    • Speech. A subtitle preset from caption styles.
    • Authored CLI copy. Overlay, with courier if the product is a prompt.

    Lock both before you generate openings. If the type is wrong in the first second, the hook line does not matter.

    Openings are a pack, not a restyle

    Once the look is a preset you would ship, buy openings. Create a branded hook pack plans N distinct hooks, charges once for the set, and submits every clip as a 1080p portrait job on Vidu Q3. Each hook comes back with its line burned on as big TikTok-style text.

    That is the product boundary. The pack tests lines, not look. TikTok-style type is how a muted feed scores a first three seconds. It is not how you ship the terminal aesthetic. Ink and Courier live on the body you already trust — the demo, the install — not inside the pack renderer.

    The agent hook pack charges once for the set is the bill. Generate the pack before the video exists is the order. Score openings cold. Attach the winner to a body that already wears the preset.

    One look, many first lines

    The expensive mistake is a new "terminal world" for every hook: five restages, zero reusable chrome. The cheap mistake is skipping the pack because the demo exists: one opening you never tested.

    Brief the offer, not the aesthetic. "CLI for generating video from a script; objection is another GPU bill" is a pack brief. "Make it look like a hacker movie" throws away the Ink card you just locked. Score distinctness when the pack returns. Then put the winner on the body that already looks like your product.

    FAQ

    Can the hook pack use Courier instead of TikTok-style type?

    Not as the pack's renderer. The pack burns large TikTok-style text on purpose so you can score a muted feed. Apply Courier and Ink on the body, or overlay your own line on a keeper. Do not expect the pack to be the brand kit.

    Is Ink the "terminal" preset?

    There is no preset named terminal. Ink is the black-card family. Courier is the mono in the font registry. Combine them. That combination is the look. A generate of "green phosphor CRT" is a different product.

    Do I need a hook pack if the product is a screenshot?

    If you are posting, you need a first three seconds. A screenshot is a still. A pack is openings against a body. Style the screenshot's type in caption styles, then test lines. Skip the pack only if you are not testing.

    Why not caption the demo and call that the hook?

    Captions of speech are a transcript. A hook is a claim the user did not say. Overlay the claim. Preset the look. Then, if you need variants of the claim, buy the pack.