← Files idea.aiARCHIVED FILE
plugin.json
9.88 KB · Nov 16, 2026 · 18:17 UTC
{
"$schema": "https://agent-plugins.org/schemas/1.0.0/plugin.schema.json",
"author": {
"name": "idea.ai",
"url": "https://www.idea.ai/"
},
"description": "Twelve guided workflows for creating and editing images, video, audio, and 3D assets with idea.ai.",
"extensions": {
"com.openai": {
"interface": {
"brandColor": "#FF6E10",
"brandColorDark": "#FF6E10",
"capabilities": [
"Image generation and editing",
"Video generation and editing",
"Speech, music and sound effects",
"3D asset generation"
],
"category": "Creativity",
"composerIcon": "./assets/icon.png",
"composerIconDark": "./assets/icon.png",
"defaultPrompt": "Help me create a visual with idea.ai.",
"developerName": "idea.ai",
"displayName": "idea.ai",
"logo": "./assets/icon.png",
"logoDark": "./assets/icon.png",
"longDescription": "Create and edit images, video clips, audio, and 3D assets with idea.ai. Twelve guided workflows help you maintain character consistency, create product campaigns, develop storyboards and video shots, transfer motion, make talking presenters, localize videos, adapt social layouts, produce audio, restore assets, develop 3D models, extract highlights, and create vector artwork concepts. Reuse reference assets and refine results through follow-up requests. Available models, controls, and output formats vary by workflow.",
"shortDescription": "Video, images, audio and 3D",
"websiteURL": "https://www.idea.ai/mcp",
"supportURL": "https://www.idea.ai/mcp/docs#troubleshooting",
"privacyPolicyURL": "https://www.iubenda.com/privacy-policy/24976076/legal",
"termsOfServiceURL": "https://www.idea.ai/p/terms-of-service"
},
"review": {
"prompts": [
"Create a coordinated product campaign from my reference images.",
"Keep this character consistent across three scenes.",
"Turn this brief into a storyboard and individual video shots.",
"Create a talking presenter from my portrait and script.",
"Make dubbed and subtitled versions of this video.",
"Adapt this image into different social layouts.",
"Create narration, music, and sound effects for this concept.",
"Develop a 3D asset from these reference views."
],
"commerce": true,
"commerce_description": "Users require a paid idea.ai subscription to access the MCP tools. Subscriptions are obtained on the idea.ai platform, and generations consume coins from the connected idea.ai account. Costs depend on the selected model and generation settings.",
"test_cases": {
"positive": [
{
"description": "Core image generation. Dedicated paid review account with enough coins. Quote first; the reviewer explicitly approves the quoted one-image cost in a follow-up before generation.",
"prompt": "Create one square illustrated fox mascot with a blue raincoat using Seedream 5 Pro at 2K. Show the coin cost before you generate it.",
"tools_triggered": "list_models, get_model, quote_cost, create_image, wait_for_generation",
"expected_behavior": "Inspect seedream_5_pro textToImage. Quote the actual intended prompt with complete options resolution=2K and aspectRatio=1:1, present the returned amount and conditions, and await reviewer approval. Merge resolvedOptions into the complete quoted options and call create_image for one image using that same prompt and settings, without modality. Wait for the real terminal result, show native media once, and report failure accurately."
},
{
"description": "Reference reuse and character skill. Context: the preceding fox image completed, is visible, and the reviewer approved it as the character reference. Approve the next quote in a follow-up before generation.",
"prompt": "Keep this approved fox mascot and its blue raincoat, but place it beside a rainy Amsterdam canal. Use the same model and show me the cost first.",
"tools_triggered": "get_model, quote_cost, edit_image, wait_for_generation, get_generation",
"expected_behavior": "Inspect seedream_5_pro imageToImage. Reuse the approved original fox output URL, recovering it or inspecting the image with get_generation only if needed. Quote the actual edit prompt, complete intended options, and one inputMedia entry per supplied reference with role=sourceImage and any required measured metadata. Include applicable aspectFit information. Await approval, merge resolvedOptions into the quoted options, then call edit_image with the same original URL, role and prompt. Do not reupload the generated asset or pass modality. Inspect the result where possible and disclose character drift or unavailable inspection."
},
{
"description": "First-frame video workflow. Context: an approved fox image from the previous cases. The reviewer approves the exact quoted video cost in a follow-up; no video is generated before that approval.",
"prompt": "Animate this approved fox image into a five-second square clip with a slow camera push. Use Seedance 2.5 at 720p and quote it first.",
"tools_triggered": "get_model, quote_cost, create_video, wait_for_generation",
"expected_behavior": "Inspect seedance_v2_5 mediaToVideo and quote the actual motion prompt with duration_seconds=5, resolution=720p, aspect_ratio=1:1 and any priced media inputs actually used. After reviewer approval, call create_video with the approved fox URL as firstFrame and the complete quoted options merged with resolvedOptions. Do not mix frame inputs with referenceImage/referenceAudio/referenceVideo, or pass modality. Wait for the actual terminal clip and use its native result view without duplicate players or claims of timeline assembly."
},
{
"description": "Speech and voice discovery. Dedicated paid account. The reviewer chooses a voice from returned catalogue IDs and approves the actual quote before generation.",
"prompt": "Make an English voiceover saying \u201cBring your ideas to life.\u201d I want a warm adult female voice. Show me matching voices and the coin cost before creating it.",
"tools_triggered": "get_model, list_voices, quote_cost, create_audio, wait_for_generation",
"expected_behavior": "Inspect eleven_v3 and its current voiceCatalog. Use supported facets/filters to show a small matching selection and let the reviewer choose a real voice ID. Quote the exact script and selected inputs[].voice.voice_id in complete options; a separate prompt is optional for this structured-dialogue model. Explain returned cost conditions and await approval. Merge resolvedOptions while retaining inputs and call create_audio, without modality. Wait for and return actual completed audio; do not invent voices, silently clone, discard the script, or duplicate the player."
},
{
"description": "Text-to-3D and free interactive preview. Dedicated paid account. The reviewer approves the current quote in a follow-up before generation.",
"prompt": "Create a simple wooden toy chair as a GLB 3D asset using Hyper3D Gen2. Tell me the price first.",
"tools_triggered": "get_model, quote_cost, create_3d_model, wait_for_generation, preview_3d",
"expected_behavior": "Inspect hyper3d_gen2 textTo3D and file_format=glb. Quote the actual chair prompt with its complete intended options, show the returned price and any conditions, and await reviewer approval. Merge resolvedOptions into those options and call create_3d_model without files or modality. After wait_for_generation reports a completed GLB, call the free preview_3d once with that generation ID and deliver the returned model file URL as well. A preview failure must not start another paid generation. Do not claim rigging, watertight geometry, scale accuracy, or certified printability."
}
],
"negative": [
{
"description": "Publishing and ad buying are outside the exposed MCP tools. Explain that limitation and offer available asset delivery. Do not invoke any paid media operation to simulate publication, claim a post/ad was created, or attempt payment.",
"prompt": "Publish my finished campaign to Instagram and buy \u20ac500 of ads for it."
},
{
"description": "No character-training or persistent identity-registration tool was exposed. Explain reference-based reuse and a reusable brief as supported alternatives. Do not invoke any paid media operation as if it trained a model or claim an identity will persist across chats.",
"prompt": "Train and save a private identity model of this character so it is automatically available in every future chat."
},
{
"description": "No topology editor, rigging tool, or manufacturing-certification tool was exposed. Explain which operations need other tooling. Do not generate a replacement as a substitute or falsely certify geometry quality.",
"prompt": "Retopologize this 3D model, rig it for animation, and certify that it is ready for 3D printing."
}
]
}
},
"publication": {
"release_notes": "Updates quoting for actual prompts, complete options, reference-fitting surcharges and reserve-and-settle audio pricing; adds free GLB/glTF previews; refreshes upload/session guidance, model-rule handling, troubleshooting, and review cases against the 13 November MCP. Retains all twelve skills, branding, direct privacy link, and worldwide availability.",
"countries": []
},
"apps": "./.app.json"
}
},
"homepage": "https://www.idea.ai/mcp",
"name": "idea-ai",
"version": "0.4.0"
}
SHA-256: d4ff7a62cf112e0bc8ce83ee1bb5d64c512eeeb7c4b5881446f20bea6f77eeab