---
title: "Reel cover set — Meleon Skill"
description: "Make a matching set of 6 Reels and TikTok covers for one series, so a profile grid reads as a set. One art direction for every cover, rendered text-free with GPT Image 2 (cheap low drafts, then medium finals, with cover 1 as the look reference for the rest), then each exact title drawn locally so every word is spelled right. A check script fails any title that falls outside the 3:4 profile-grid crop or under the Reels and TikTok buttons and captions, any title too small to read on a grid tile, and low contrast. A grid script shows the set as followers see it, 3 x 2, newest first. No face needed; the creator's own video stills can be restyled into the look, with consent and likeness rules for any person in them. For creators, social teams, brands and agencies."
url: https://meleon.ai/skills/reel-cover-set/
author: "Meleon AI"
---

# Reel cover set

> Make a matching set of 6 Reels and TikTok covers for a series, each with an exact title that stays readable on the profile grid.

- Type: Image Skill
- Category: Social posts & covers
- Version: 1.0.0 (updated 2026-11-09)
- Made by: Meleon AI (official)
- Price: Free to download. Outside costs: About $0.74 for a set of 6 on fal (GPT Image 2, low drafts about $0.03, medium finals about $0.71), up to about $0.99 with two re-rolls; about $0.12 a re-roll. The titles, the grid preview and the checks are free. Billed to your own fal account or your Meleon credits; quote the total and get a yes before the first paid call.
- Runs in: Claude Code, Codex and Cursor
- Page: https://meleon.ai/skills/reel-cover-set/
- Download: https://meleon.ai/downloads/reel-cover-set.zip
- Manifest: https://meleon.ai/skills/reel-cover-set.json

## What it does

Make a matching set of 6 Reels and TikTok covers for one series, so a profile grid reads as a set. One art direction for every cover, rendered text-free with GPT Image 2 (cheap low drafts, then medium finals, with cover 1 as the look reference for the rest), then each exact title drawn locally so every word is spelled right. A check script fails any title that falls outside the 3:4 profile-grid crop or under the Reels and TikTok buttons and captions, any title too small to read on a grid tile, and low contrast. A grid script shows the set as followers see it, 3 x 2, newest first. No face needed; the creator's own video stills can be restyled into the look, with consent and likeness rules for any person in them. For creators, social teams, brands and agencies.

## Made for

- Creators
- Social teams
- Brands
- Agencies

## Before you start

- You give it: Series name + 6 titles; own video stills optional
- You need: Node.js 18 or newer for the scripts; the sharp package for the title and grid steps (run `npm install sharp` in the Skill folder, Node.js 20.9 or newer); your own FAL_KEY only if you run scripts/fal-run.mjs.
- Cost per run: About $0.74 for a set of 6 on fal (GPT Image 2, low drafts about $0.03, medium finals about $0.71), up to about $0.99 with two re-rolls; about $0.12 a re-roll. The titles, the grid preview and the checks are free. Billed to your own fal account or your Meleon credits; quote the total and get a yes before the first paid call.
- Makes: 6 PNG covers, 9:16 (1080 x 1920), + a 3 x 2 profile-grid preview PNG

## Install

Run the command for your agent in a terminal. It downloads https://meleon.ai/downloads/reel-cover-set.zip and unzips it into the agent's personal Skills folder.

### Claude Code

```sh
mkdir -p ~/.claude/skills && curl -fsSL https://meleon.ai/downloads/reel-cover-set.zip -o /tmp/reel-cover-set.zip && unzip -o /tmp/reel-cover-set.zip -d ~/.claude/skills
```

Then ask Claude Code to use it: `/reel-cover-set`.

### Codex

```sh
mkdir -p ~/.agents/skills && curl -fsSL https://meleon.ai/downloads/reel-cover-set.zip -o /tmp/reel-cover-set.zip && unzip -o /tmp/reel-cover-set.zip -d ~/.agents/skills
```

Then ask Codex to use it: `$reel-cover-set`.

### Cursor

```sh
mkdir -p ~/.cursor/skills && curl -fsSL https://meleon.ai/downloads/reel-cover-set.zip -o /tmp/reel-cover-set.zip && unzip -o /tmp/reel-cover-set.zip -d ~/.cursor/skills
```

Then ask Cursor to use it: `/reel-cover-set`.

## Example request

```text
/reel-cover-set my recipe series "10-Minute Dinners", six titles attached, bold flat look
```

In Codex, start with `$reel-cover-set` instead of `/reel-cover-set`.

## Files

Every file in the package, readable before you install:

- [SKILL.md](https://meleon.ai/skills/reel-cover-set/files/SKILL.md) (11.9 KB)
- [references/cover-grammar.md](https://meleon.ai/skills/reel-cover-set/files/references/cover-grammar.md) (5.7 KB)
- [references/output-spec.md](https://meleon.ai/skills/reel-cover-set/files/references/output-spec.md) (5.3 KB)
- [references/prompt-patterns.md](https://meleon.ai/skills/reel-cover-set/files/references/prompt-patterns.md) (6.8 KB)
- [assets/generation-input.json](https://meleon.ai/skills/reel-cover-set/files/assets/generation-input.json) (5.3 KB)
- [assets/preview.webp](https://meleon.ai/skills/reel-cover-set/files/assets/preview.webp/) (158.9 KB)
- [assets/prompt-template.md](https://meleon.ai/skills/reel-cover-set/files/assets/prompt-template.md) (390 bytes)
- [assets/safe-zones.json](https://meleon.ai/skills/reel-cover-set/files/assets/safe-zones.json) (3.1 KB)
- [assets/series.example.json](https://meleon.ai/skills/reel-cover-set/files/assets/series.example.json) (2.7 KB)
- [scripts/build-requests.mjs](https://meleon.ai/skills/reel-cover-set/files/scripts/build-requests.mjs) (14.2 KB)
- [scripts/check-covers.mjs](https://meleon.ai/skills/reel-cover-set/files/scripts/check-covers.mjs) (13.9 KB)
- [scripts/compose-covers.mjs](https://meleon.ai/skills/reel-cover-set/files/scripts/compose-covers.mjs) (17.5 KB)
- [scripts/fal-run.mjs](https://meleon.ai/skills/reel-cover-set/files/scripts/fal-run.mjs) (8.2 KB)
- [scripts/grid-preview.mjs](https://meleon.ai/skills/reel-cover-set/files/scripts/grid-preview.mjs) (5.2 KB)
- [scripts/new-series.mjs](https://meleon.ai/skills/reel-cover-set/files/scripts/new-series.mjs) (7.6 KB)

## Versions

- 1.0.0 (2026-11-09)

## Instructions (SKILL.md)

Make six covers for a Reels or TikTok series that look like one set on the profile grid. Every cover shares one art direction, every title is spelled exactly (the image model never draws words), and every title sits where the grid and the app leave it alone: inside the 3:4 crop Instagram shows on the profile, and clear of the buttons and captions in the player. A grid preview shows the set the way followers will see it.

### What you need

- **The series name** and **6 titles**, one per episode (3 or 9 also work: full rows of the grid). Each title 1 to 6 words, at most 32 characters. Draft them with the user from what each episode really contains, and get approval. No invented results, prices, awards or claims.
- **A look**: a preset (`bold-flat`, `moody-film`, `soft-pastel`, `paper-collage`, `risograph`; see `references/cover-grammar.md`, "Series patterns") or the user's own one-sentence art direction. There is no default look, so a set never copies the sample.
- **Optional: a kicker**, a short line above each title such as the series name or `EP {{n}}`.
- **Optional: brand colours and a font file** (`.ttf` or `.otf`) for the titles.
- **Optional: stills from the user's own videos**, one frame per episode (export a JPEG from the video first). Only their own footage or footage they have rights to. Crop out other people, number plates, screens and other brands' logos.
- **People.** No face is needed, and object, food, place and graphic covers are the default. If a still shows a person, it must be the user or someone who has agreed in plain words to appear on the covers. Ask, and stop if the answer is not a clear yes. Never a celebrity, a public figure or a child, and never a photo of someone who has not agreed. The person stays exactly themselves: same identity, age range, skin tone, hair, features and body. Make no cover that suggests an endorsement, a result or a reaction that did not happen. Generated people are fictional adults who resemble no real or famous person.

If the titles or the look are missing, ask for them.

### What you make

1. `covers/cover-01.png` ... `cover-06.png`: 1080 x 1920 PNG covers with the exact title (and kicker) drawn locally, all in one title size, place and style. Upload each as the custom cover of its Reel or TikTok.
2. `grid.png`: the set as a 3 x 2 profile grid of 3:4 tiles, newest first (cover 6 top left), and an optional `preview.webp` for a message.
3. `covers/manifest.json` and `series.json`: the title boxes and the plan, so titles can be redrawn for free.
4. `art/`: the text-free drafts and finals.

### How to run it

**Budget: about $0.74 for a set of 6 (6 low drafts about $0.03, 6 medium finals about $0.71), up to about $0.99 with two re-rolls, about $0.12 a re-roll. A still adds about $0.01 a call. Titles, grid and checks are free.**

Before the first paid call, run `node scripts/build-requests.mjs --series series.json --stage estimate`, quote the total it prints, and get one explicit yes on it. At every paid step, tell the user the model, the number of calls and the cost, and ask again if the spend will pass the quoted total.

To send a request: if the Meleon tools are connected, call `generate_image` for a request without `image_urls`, or `edit_image` for one with them: its `prompt`, its `image_urls` (write the requests with `--out requests --inline`, so local images are data:image/ URIs; Meleon deletes uploads after 7 days), `model` "gpt-image-2", its `image_size` as `size`, its `quality` and `output_format`, `count` 1, and `skill` "reel-cover-set", so the run shows under this Skill in the user's Meleon dashboard. It costs Meleon credits (the price is in the tool's description), or nothing if the user saved their own fal key on Meleon. Otherwise POST the printed `input` to the fal model id in the request with the user's own FAL_KEY (`node scripts/fal-run.mjs --request <file> --out art --approved` does this; the Skill never stores keys), or paste the prompt and images into any GPT-Image-class editor.

1. **Brief (free).** Confirm the series name, the titles (letter for letter), the look, the kicker, colours and font, any stills (and the rights and people questions above), and the title treatment and position (`references/cover-grammar.md`).
2. **Plan (free).** `node scripts/new-series.mjs --out series.json --name "..." --titles "t1|t2|t3|t4|t5|t6" (--preset <name> | --look "..." [--palette "colour 1|colour 2"]) [--kicker "EP {{n}}"] [--treatment block|shadow|plain] [--position top|middle|bottom] [--align left|center] [--case as-written|upper] [--ink #hex --accent #hex] [--font-file brand.ttf --font "Family"] [--stills "ep1.jpg||ep3.jpg"] [--people none|fictional|real-with-consent]`. Then write each cover's `subject` in `series.json` (`references/prompt-patterns.md`): the hero thing for that episode, the same kind and size on every cover, no words in it. For a still with a person, set its `still_has_person` to true, `series.people` to `real-with-consent`, and `series.consent_confirmed` to true only after the yes. Show the user the plan in plain words.
3. **Spend gate (free).** `node scripts/build-requests.mjs --series series.json --stage estimate`. Quote the total and wait for one explicit yes.
4. **Drafts (about $0.03 for six).** `node scripts/build-requests.mjs --series series.json --stage draft --out requests --inline`, then send each `requests/cover-0N-draft.json` and save the results in `art/`. Lay the six side by side and check: one look, the subject in the same place and size, a calm empty band where the title goes, no stray text. Re-roll a draft by changing one thing in its subject. A long title that wraps to three lines needs more room: say the subject is small and low in that cover's `subject` ("its top edge no higher than halfway down the image"). Put each approved draft's path in that cover's `draft`.
5. **Cover 1 final (about $0.11).** `--stage final --cover 1 --out requests --inline`, send it, open it at full size. When it is right, put its path in cover 1's `art` and in `references.anchor`.
6. **The other finals (about $0.12 each).** `--stage final --cover rest --out requests --inline`. Each carries cover 1 as the look reference and its own draft for the composition. Open every final and check it against cover 1 and `references/output-spec.md`. At most 3 attempts per cover, one change each. If a re-roll changes the composition, drop that cover's draft from `image_urls` (and its sentence from the prompt), or the draft pulls the old layout back. Put each approved path in that cover's `art`.
7. **Titles (free).** Run `npm install sharp` in the Skill folder once, then `node scripts/compose-covers.mjs --series series.json --out-dir covers --guides`. It draws every title and kicker inside the safe title box, at one size for the whole set (the largest every title fits at, 96 to 156 px), and writes `covers/manifest.json`. `--guides` writes copies with the zones drawn over them. Fonts: Helvetica Neue Bold by default (DejaVu on Linux); pass `--font-file brand.ttf --font "Family Name"` for brand type. To change a title, a colour or the treatment, edit `series.json` and run it again: no new art needed.
8. **Grid (free).** `node scripts/grid-preview.mjs --manifest covers/manifest.json --out grid.png --webp preview.webp`. Newest first by default: post the covers as episode 1 to 6 and the grid reads 6, 5, 4 over 3, 2, 1.
9. **Check (free).** `node scripts/check-covers.mjs --series series.json --manifest covers/manifest.json --grid grid.png --preview preview.webp`. It fails a title outside the 3:4 grid crop or in a Reels or TikTok interface zone, a title that does not match the plan, a title under 96 px, contrast under 3:1, a title or kicker block that sits on the art (more than 1 percent of the pixels around it are not backdrop: a dish, a prop or a hard shadow under or touching it), the wrong size, missing consent, or files out of date. Then open every cover, the guides and the grid at full size and go through `references/output-spec.md` by eye, reading every title letter by letter. Fix and recompose until it passes.
10. **Deliver.** The six covers in order, the grid and preview, `series.json` and the art. Say which covers were re-rolled and anything still imperfect, and give the posting order.
11. **Clean up photos.** Delete the user's stills from the working folders unless they want them kept. fal keeps request inputs and outputs on its side for a period: tell the user they can delete them from their fal dashboard, and to check fal's current retention policy.

### Files

- `references/cover-grammar.md`: where covers are seen, the safe title box and its zones, title rules, treatments, what makes a set read as one, posting order, series patterns.
- `references/prompt-patterns.md`: prompt anatomy, writing the look and the subjects, stills, people and consent, drafts and finals, failure modes and fixes.
- `references/output-spec.md`: deliverables, sizes, costs, the pass/fail checklist.
- `assets/prompt-template.md`: the cover prompt with placeholders.
- `assets/generation-input.json`: model ids, draft and final sizes and qualities, approximate costs, the composition bands, the look presets.
- `assets/safe-zones.json`: the frame, the safe title box, every zone with its source, the grid crop, the minimum title size and contrast.
- `assets/series.example.json`: the sample plan (10-Minute Dinners, an invented recipe series, `bold-flat` with a tomato red and butter yellow checkerboard).
- `scripts/new-series.mjs`: writes series.json and checks the titles.
- `scripts/build-requests.mjs`: the estimate and the draft and final requests. Calls no API.
- `scripts/fal-run.mjs`: sends one request with the user's FAL_KEY after `--approved`, waits, downloads (never overwrites without `--force`).
- `scripts/compose-covers.mjs`: draws the exact titles and kickers, writes the manifest and the zone guides (needs sharp).
- `scripts/grid-preview.mjs`: the 3 x 2 profile-grid preview (needs sharp).
- `scripts/check-covers.mjs`: safe zones, exact titles, size, contrast, consent and freshness checks.

## Related

- [Viral video thumbnail](https://meleon.ai/skills/viral-video-thumbnail.md): Turn a portrait into a bold video thumbnail.
- [Film look grade](https://meleon.ai/skills/film-look-grade.md): Give any photo a real film look, graded locally in seconds.
- [Open-world loading screen](https://meleon.ai/skills/open-world-loading-screen.md): Turn your pet, product or character into a cocky open-world loading screen with an exact caption and a loading bar.
- [All Meleon Skills](https://meleon.ai/skills/) (index as JSON: https://meleon.ai/skills.json; site guide: https://meleon.ai/llms.txt)
