ClawKit Logo
ClawKitReliability Toolkit
Back to Registry
Official Verified

agentboard

Build multi-panel storyboards programmatically — create projects, upload images/audio to boards, composite annotations, export PDFs, share via public URL. Invoke when the user wants a storyboard, pre-visualization, shot breakdown, animatic, or any ordered sequence of visual panels with text. Hosted at https://agentboard.fly.dev. Works over REST from any agent; MCP tools (mcp__agentboard__*) available in compatible runtimes. If you have your own image/audio generator, use it and UPLOAD the bytes — optional server-side generation endpoints exist as a fallback only for agents without built-in generation.

skill-install — Terminal

Install via CLI (Recommended)

clawhub install openclaw/skills/skills/0xartex/agentboard
Or

AgentBoard

Base URL: https://agentboard.fly.dev. All examples below use REST. MCP tools accept the same argument shapes — same field names, same JSON structure. No client credentials are required to call the hosted endpoints.

Data model

  • Project — UUID, aspectRatio (default 1.7777), ordered list of boards.
  • Board — 5-char uid, dialogue, action, notes, duration (ms), up to 6 layers (fill is primary; others: tone, pencil, ink, reference, notes).
  • Layer asset — one image per (board, layer). Re-upload replaces.
  • Audio asset — one audio per (board, kind). Kinds: narration, sfx, music, ambient, reference. Multiple kinds per board allowed.

Workflow

1. create project + boards (one call)
2. populate: upload OR generate images/audio
3. (optional) annotate via draw_shapes
4. (optional) export PDF and/or get share URL

1. Create project

POST /api/agent/create-project
{
  "title": "The Lighthouse Keeper",
  "aspectRatio": 1.7777,
  "boards": [
    { "dialogue": "...", "action": "...", "duration": 4000 },
    { "dialogue": "...", "action": "...", "duration": 3500 }
  ]
}
→ { id, project: { boards: [{ uid, number, ... }] }, viewUrl }

Save id and every board's uid from the response. Rate-limited: 50 creates/hour per IP+token.

2. Populate boards — upload your own bytes (PREFERRED)

Use this path if your runtime has image/audio generation (fal, Sora, Veo, Midjourney, Gemini, etc.). Generate with your tool, then upload. Skip to the Fallback generation section only if you can't generate.

Batch upload (3+ panels)

POST /api/agent/upload-batch
{
  "projectId": "...",
  "uploads": [
    { "boardUid": "ABC", "kind": "image", "imageBase64": "..." },
    { "boardUid": "DEF", "kind": "image", "imageBase64": "..." },
    { "boardUid": "ABC", "kind": "audio", "audioKind": "narration", "audioBase64": "..." }
  ]
}
→ 201 (all ok) or 207 Multi-Status { succeeded: [...], failed: [{ index, error }] }

Up to 100 items per call, mixing images and audio freely. Partial failures don't abort the batch.

MCP upload_assets_batch supports three input modes per item. Always prefer path/url over base64 — they cost zero agent context:

FieldWhatUse for
imagePath / audioPathLocal file pathPreferred when the user has explicitly produced or pointed at a file. MCP subprocess reads and uploads the bytes. Never reference paths the user didn't authorize.
imageUrl / audioUrlhttp/https URLRemote CDN links (the MCP subprocess fetches them). Zero agent context cost.
imageBase64 / audioBase64Inline base64Only for tiny images (<10 KB). Burns agent context at 1.33× file size.

Single upload

POST /api/agent/draw
{ "projectId":"...", "boardUid":"...", "layer":"fill", "imageBase64":"...", "mime":"image/png" }

Metadata

Author@0xartex
Stars4473
Views0
Updated2026-05-01
View Author Profile
AI Skill Finder

Not sure this is the right skill?

Describe what you want to build — we'll match you to the best skill from 16,000+ options.

Find the right skill
Add to Configuration

Paste this into your clawhub.json to enable this plugin.

{
  "plugins": {
    "official-0xartex-agentboard": {
      "enabled": true,
      "auto_update": true
    }
  }
}
Safety NoteClawKit audits metadata but not runtime behavior. Use with caution.