agentleFS
vargHQvargHQ/skills/varg-occasion-video/SKILL.md

varg-occasion-video

Skill23 starsChanged 8 months agoMITView on GitHub

Turns a few photos and the details of an occasion into an AI video on varg: a wedding or birthday invitation and save-the-date that unfolds (the names and date printed right), a Christmas, New Year, Diwali, Eid or Lunar New Year greeting, old family photos brought to life, a love story across the eras for an anniversary, and a calm memorial or tribute film. Asks what the occasion needs, makes the stills first (GPT Image 2.5 prints names and dates), animates them (Seedance 2.5, Kling 4, Wan 3), adds music and exports for WhatsApp, a party screen or a service. Use when someone asks for a video invitation, an e-invite, a save-the-date, a wedding invitation video for WhatsApp, a holiday or festival greeting video, an animated e-card or Christmas card, "bring old photos to life", an anniversary or love-story video, or a memorial, celebration-of-life, retirement or farewell tribute. Not for birthdays (use varg-birthday-video) or editing real wedding footage (varg-video-editing).

  • Reads credentials
  • Sends data out

The file

We split this skill into the rules it gives an agent

2 rules. About 0 are common, 2 are distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈2

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Every name, date and place matches what the person gave you, letter by letter, in the last frame.

    line 84
  2. Distinctive:

    Never a celebrity, a public figure or a stranger.

    line 106

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Invitations, holiday greetings, old photos and tributes

The occasion sets the rules: an invitation must carry the names, the date and the place without a single wrong letter; a holiday greeting must be forwardable to a family group; old photos must still look like the people; a tribute must stay calm. Generation runs on varg: REST https://api.varg.ai/v2 (the varg-ai skill) or the MCP server https://mcp.varg.ai/mcp (generate_image, generate_video, generate_music, call_tool, get_job, estimate_cost). Quote the price before anything runs.

1. Which occasion

OccasionWhat people want (searches, Oct 2026)FormatReference
Invitation / save-the-date"ai wedding invitation video", "save the date video", "invitation video for whatsapp" (851 suggestions)A card or envelope that opens onto the names, date and place; 15–30 s, 9:16references/invitations.md
Holiday greeting"ai christmas video", "christmas card video", "ai new year video", Diwali, Eid (386)An animated card, a snow globe, a lamp-lit story, a family scene; 10–30 sreferences/holidays.md
Old photos brought to life"animate old photos", "bring old photos to life" (222)Restore, then a gentle motion per photo, cut to musicreferences/old-photos.md
Anniversary / love story"ai anniversary video", "ai wedding video" (339)The couple across the eras, ending where they metreferences/old-photos.md
Memorial / tribute"memorial video ai", "celebration of life video", "retirement tribute video" (254)Restored photos, slow motion, words from the familyreferences/old-photos.md

2. Ask, then go

  • Exact words: names as they are written, the date, the time, the place, the RSVP line, the greeting. Read them back to the person before you generate anything that prints them.
  • Photos: the couple or the family (with their agreement), old prints scanned flat, or none (an illustrated card needs no face).
  • Style: pick one look (references/styles.md): pop-up card, envelope, illustrated card, snow globe, diya glow, 1990s film, cozy holiday still, retro home video.
  • Where it goes: WhatsApp or iMessage (9:16, under 60 s, send as a document to keep quality), Instagram story (9:16), a screen at the venue or the service (16:9).
  • Music: an original track (eleven_music_v2_5 for a song with words, mureka_9_5 for an instrumental) or a public-domain piece; never a commercial song on a public post.

3. Stills first, then motion

  1. Print the words on a still with gpt_image_2_5: quote every word and spell it ("the words "ANNA & LEO" and below "12 JUNE 2027 · LISBON", spelled exactly like that"). Read the still at full size; if one letter is wrong, regenerate. For anything longer than a line or two, leave the space empty and add real type later (HyperFrames or an editor): generated paragraphs smudge.
  2. Make the matching first frame as an edit of that still (generate_image with reference_image): the card closed, the envelope sealed, the photo still faded. Same table, light and angle.
  3. Animate between them with seedance_2_5 first and last frames: files: [first, last] and provider_options: {piapi: {input: {mode: "first_last_frames"}}}. The model moves from the closed card to the open one, so the words land exactly as printed. For a single photo, image-to-video on kling_v4 or wan_3.
  4. Music and export with scripts/assemble.sh (scales, joins, lays the track, -14 LUFS, H.264 + AAC).
curl -s -X POST https://api.varg.ai/v2/video -H "Authorization: Bearer $VARG_API_KEY" -H "Content-Type: application/json" \
  -d '{"model":"seedance_2_5","duration":8,"aspect_ratio":"9:16","resolution":"720p","audio":true,
       "files":[{"url":"CLOSED_CARD_URL"},{"url":"OPEN_CARD_URL"}],
       "provider_options":{"piapi":{"input":{"mode":"first_last_frames"}}},
       "prompt":"<timed beats: the card opens, the paper scene rises, the names settle; paper sounds>"}'
bash scripts/assemble.sh -s 0 invite.mp4 music.mp3 clip.mp4

4. Check before you send

  • Every name, date and place matches what the person gave you, letter by letter, in the last frame.
  • Faces still look like the people (eyes, smile lines, glasses); old photos keep their grain and era.
  • Nothing extra appears: no new guests, no rings or logos, no extra pages on a card.
  • A festival greeting shows the festival respectfully: lamps, food, family, lights; no deities unless the family asks for them.

Tested

2026-10-30, a fictional couple "Anna & Leo": the open pop-up invitation on gpt_image_2_5, 9:16 (22 credits, the names and "12 JUNE 2027 · LISBON" spelled right on the first try), the closed card as an edit of it (23 credits), then seedance_2_5 first and last frames, 8 s, 720p, with sound (280 credits): the cover swings open, the paper arch rises layer by layer, and the names and date land sharp and unchanged in the last frame, with paper sounds and a music-box chime. https://s3.varg.ai/files/72c55813-c0fa-4cf2-9f64-ec4f7ed0e7ae/file_e1ac8b0f8d63.mp4

Old photo, a fictional 1962 seaside couple: nano_banana_2_1 restore and colorize as an edit (5 credits, both faces kept), then 5 s of gentle motion on kling_v4 (79 credits: she turns to him and laughs, he smiles down at her, a slow push-in, the faces and the 1960s color and grain kept). https://s3.varg.ai/files/72c55813-c0fa-4cf2-9f64-ec4f7ed0e7ae/file_34e07d214ef1.mp4

Rules

  • Consent: animate a living person's face, or a family's old photos, only with their or the family's agreement. Never a celebrity, a public figure or a stranger.
  • The dead: no party effects, no making them speak or sing unless the family asks; keep the film private unless they want it shared.
  • Children: use the family's own photos only, deliver privately, never post a child's face publicly.
  • Label realistic AI footage as AI when it is posted on social media.

Related skills

varg-birthday-video (birthdays, the song with the name), varg-consistent-characters (the same faces in every shot), hyperframes (real type, envelopes and title cards rendered from HTML), varg-video-editing (real wedding footage), varg-ai (models and prices).

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.