A canvas for AI prompts → media → video, on your Solid pod. One self-contained HTML file, no build, signs in with xlogin.
You give it a JSON-LD document — inline as a <script type="application/ld+json">
data island, or a pod file via ?src=<url> — and it renders it as a node canvas:
prompts, image/video assets, and the edges between them (what was made from what).
The whole thing is your data on your pod, so "fork" = copy the doc to your pod —
a genuinely decentralized remix, no SaaS, no credits wall.
Inspired by the prompt-canvas pattern (promptsref, ComfyUI, Krea, Flora). The video-assembly side aligns with OpenTimeline Lite (Timeline → Tracks → Clips) so a webprompts doc can describe a full sequence, not just a moodboard. Always intended for prompt sharing (webprompts.org).
- Canvas reader — load a webprompts JSON-LD (data island or
?src=), render prompt / image / video nodes + lineage edges on a pan-zoom canvas. Read-only. - Edit + save to pod — add / move / connect / edit nodes; save to
/public/webprompts/<name>.jsonld; fork = copy to your pod. - Timeline (OTIO-lite) — arrange clips into tracks, preview in order, import/export OTIO-lite — assemble a video sequence.
- Generation — per-node Generate via a BYO model endpoint or a NIP-90 nostr DVM; output saved to your pod, becomes a new node/clip.
- Prompt library / sharing — browse, tag, and fork prompts.
One JSON-LD doc combining schema.org (media: ImageObject / VideoObject) + a
small wp: vocabulary (prompt text, model + params, canvas x/y,
derivedFrom edges) + OTIO-lite for the timeline. A prompt canvas and a video
timeline are two views of the same document.
Static — open index.html, or install via the store to /public/apps/webprompts/.
AGPL-3.0-only.