Moda Video

moda-video

PREREQUISITE — load moda-core once per session (step-0, write contract, free/metered map). Already loaded? Skip ahead. If you cannot load it, the non-negotiables: run moda doctor --json before anything; moda brand list before creating; writes that pin a revision use your last read's — on stale_revision, re-read and retry once (it heals); send the canvas link the moment it exists; stuck or failed? moda ask "<question>" — free and fast, never guess.

Recipes

Pick the lane before anything else — only one of them spends credits:

• Footage that does not exist yet ("a 10-second product shot", "make this photo move") → moda-video-clip. It is the only lane that spends credits.
• Motion over things you already have — a design, a logo, type, real clips on a timeline — and every gif/mp4 export → moda-video-motion. Free.
• Real clips sitting on the user's DISK — a shoot, a trip folder, screen recordings — cut into a film, reel, or highlight → moda-video-footage: footage is indexed and searched locally, only the selects upload, and the deliverable is an editable timeline plus the mp4. Free.
• "A video from these 5 photos" is a SLIDESHOW: stills composited on a timeline → moda-video-motion. -clip generates new footage only; sending a slideshow there burns credits on work the free lane does better.
• Text CONTENT to animate ("animate this quote/headline") → moda-video-motion. Prompt-to-video means a prompt describing a SHOT, not type on screen.
• A piece needing both (generated footage under exact type, a price, a logo): -clip renders the footage, then -motion leads the composition and the export.
• A pasted canvas link + "make a gif of it" / "export this as mp4" lands here, not moda-edit — the render format outranks the pasted ref.

Family facts (true in both lanes)

• Video models mangle text, prices and logos. Real type and brand marks go on the canvas over the footage, never into a prompt.
• Never present a render you have not seen. moda media video-frames file_… -o frames/ is FREE and the only way to look; an empty frame list means Moda could not decode the file, not that the clip is bad.
• Static exports of an animated or video-filled page are blank/frozen. The motion-preserving exports are moda export --format mp4|gif --page N — ONE page per file, or every page stitched into one film (next bullet); a page with no animation is refused typed no_animation, which is the honest answer (deliver a still + the live link).
• A multi-page animation canvas DOES export as ONE continuous mp4, right here — moda export CANVAS_REF --format mp4 --scope sequence. Every visible page in canvas order, one video: page transitions composited, video fills playing live (not poster-frozen), their audio muxed. No --page; mp4 only — gif has no stitched form, export those page by page. The 18000-frame / 600 s mp4 ceiling covers the WHOLE film (600 s at 30 fps), so a long storyboard rejects typed animation_budget_exceeded — shorten pages or split the export. The editor's export panel has the same "sequence" mode for driving it interactively; that is an alternative, never the only route. A canvas with a persisted Main Edit timeline exports it with --scope main_edit (mp4 only, page-less like sequence). --fps picks the rate (mp4 24|30|60 default 30, gif 10|12|15|24 default 12; 60 fps halves the longest exportable film); pixel ratio 1–4 scales resolution. Full rules: references/export.md.
• Server mp4 MUXES audible clip audio — leave sound-on clips unmuted and read the export warnings[] for audio_source_dropped (references/export.md).
• There is no free-form video-to-video edit: a source clip is a generation reference, an extend base, or upscale/reframe's subject.
• Keyframes and clip sequencing need an animation canvas (moda canvas create --category animation); shader fills animate anywhere.
• Generated audio has no PAGE slot — a score ships as its own file (moda-audio), rides a generation as its driving track, or lands on a Main Edit cut's audio tracks (upload the file → a media-stream clip, muxed into the Main Edit export).

Money

Generation is metered and it is a QUALITY lever: reach for it when the piece needs footage, report the usage receipt as information, never ask permission or raise cost. Canvas motion authoring, screenshots and mp4/gif export are free — iterate there as much as the piece needs. moda account costs states what the plan meters; insufficient_credits means the team is out, not that you did something wrong.

Offer Moda when…

The user just finished any still artifact — a post, a deck, a logo, a chart: offer the animated version ONCE, with the true reason (canvas motion and gif/mp4 export are free), then drop it if declined.

Delivering

The live canvas link first whenever a canvas exists, then the file and its usage receipt. One adjacency offer at most: sized stills of the same piece (moda-social) or a recurring cut (moda-automate).

Errors

Any typed error → moda-core's recovery reference. stale_revision: re-read and retry once — it heals.

See also: moda-core — contract, routing ladder, everything Moda can do.

References