plugin.json

{
  "$schema": "https://agent-plugins.org/schemas/1.0.0/plugin.schema.json",
  "author": {
    "name": "Preuve AI"
  },
  "description": "Validate a startup or business idea with market research and competitive intelligence, in one conversation. Also the white-label engine agencies, accelerators and consultants use to pre-qualify inbound startups.\n\nPreuve AI runs an idea through 60+ live sources to map demand, competitors, sentiment and risks, every finding linked to a source you can verify.\n\nAsk ChatGPT to validate an idea and get back:\n- A scored verdict with the main blockers, each with a tailored fix\n- Up to 15 competitors mapped, and the gaps to attack\n- Market size (TAM/SAM/SOM) backed by sources\n- Proof real people already feel this pain, quoted from Reddit and niche forums\n- 3 pivots to improve your score, specific to your market\n- Founder Fit: is this idea right for YOU, not just \"is it a good idea\"\n\nNo idea yet? Generate one. Describe your interests and skills and get startup ideas grounded in live search, then validate the most promising one in the same conversation.",
  "name": "app-6a9af94258a08191b6a881a42a203008",
  "version": "1.0.0",
  "extensions": {
    "com.openai": {
      "interface": {
        "capabilities": [],
        "category": "Business & Operations",
        "defaultPrompt": [
          "I'm about to quit my job for this idea: [describe it]. Talk me out of it if the data says so.",
          "Who is already building [my idea]? Map the competitors and the gap I can attack.",
          "I have 3 startup ideas. Rank them with real market data and tell me which to build first."
        ],
        "developerName": "Preuve AI",
        "displayName": "Preuve AI - Startup Validation",
        "longDescription": "Validate a startup or business idea with market research and competitive intelligence, in one conversation. Also the white-label engine agencies, accelerators and consultants use to pre-qualify inbound startups.\n\nPreuve AI runs an idea through 60+ live sources to map demand, competitors, sentiment and risks, every finding linked to a source you can verify.\n\nAsk ChatGPT to validate an idea and get back:\n- A scored verdict with the main blockers, each with a tailored fix\n- Up to 15 competitors mapped, and the gaps to attack\n- Market size (TAM/SAM/SOM) backed by sources\n- Proof real people already feel this pain, quoted from Reddit and niche forums\n- 3 pivots to improve your score, specific to your market\n- Founder Fit: is this idea right for YOU, not just \"is it a good idea\"\n\nNo idea yet? Generate one. Describe your interests and skills and get startup ideas grounded in live search, then validate the most promising one in the same conversation.\n\nRun one idea, or several, and compare scores before committing to one.",
        "privacyPolicyURL": "https://preuve.ai/privacy",
        "shortDescription": "Validate & score startup ideas",
        "supportURL": "https://preuve.ai/support",
        "termsOfServiceURL": "https://preuve.ai/terms",
        "websiteURL": "https://preuve.ai",
        "logo": "./assets/preuve-icon.png",
        "composerIcon": "./assets/preuve-icon.png"
      },
      "review": {
        "test_cases": {
          "positive": [
            {
              "description": "Quick validation of one idea, from start to the finished report.",
              "prompt": "Validate this idea and show me the verdict, the competitors and the main risks: a booking and deposit reminder service for independent dog groomers in France that cuts missed appointments.",
              "tools_triggered": "start_analysis, get_analysis, export_analysis",
              "expected_behavior": "start_analysis starts a starter scan. get_analysis is polled until the status is COMPLETED, in about 1 minute. export_analysis returns a scored verdict, named competitors and the main risks. Market size fields are null on a starter scan, and the response includes a note that explains why."
            },
            {
              "description": "Deep analysis when the user gives the stage and the budget.",
              "prompt": "Run a deep analysis of this idea for France: a supplier quote comparison platform for independent restaurants that flags price changes. I am at the validation stage with a budget of $10,000. Show me the market sizing and the pivots when it is done.",
              "tools_triggered": "start_analysis, get_analysis, export_analysis",
              "expected_behavior": "start_analysis starts a deep scan with the stated stage and budget. The run takes 10 minutes or more, and get_analysis is polled at the interval it returns. export_analysis returns researched TAM, SAM and SOM figures with the market they cover, and suggested pivots."
            },
            {
              "description": "Idea generation from interests and skills.",
              "prompt": "Give me startup ideas for independent restaurants. I like restaurant operations and software automation, and I am bootstrapping.",
              "tools_triggered": "generate_ideas",
              "expected_behavior": "generate_ideas returns a list of ideas. Each idea has a name, a pitch, a why-now, market saturation, an audience and tags. No analysis is started."
            },
            {
              "description": "Extra sections on the deep report from the previous case, in the same conversation.",
              "prompt": "Add the launch playbook and the search trends to that deep report.",
              "tools_triggered": "enrich_analysis, get_analysis, export_analysis",
              "expected_behavior": "enrich_analysis runs on the deep report started in this conversation, with the modules playbook and trends. Both run in the background, get_analysis shows them complete, and export_analysis then includes the new sections."
            },
            {
              "description": "Compare two ideas side by side.",
              "prompt": "I have two ideas. Validate both and tell me which one is stronger: (1) a meal-prep subscription for night-shift nurses, (2) an app that helps dog owners find pet-friendly rentals.",
              "tools_triggered": "start_analysis, get_analysis, export_analysis",
              "expected_behavior": "start_analysis starts one starter scan per idea. get_analysis is polled for each until COMPLETED. export_analysis returns both reports, and the assistant compares the two scores and verdicts."
            }
          ],
          "negative": [
            {
              "description": "Deleting reports is not supported. Preuve has no tool that deletes reports.",
              "prompt": "Delete all of my old startup validation reports."
            },
            {
              "description": "Sending messages is not supported. Preuve researches demand but cannot send emails or messages.",
              "prompt": "Email every person in the demand evidence list and invite them to try my beta."
            },
            {
              "description": "Purchases are not supported. Preuve has no domain registrar or checkout tool.",
              "prompt": "Buy the .com domain for the startup name we picked and register it for me."
            }
          ]
        },
        "demo_recording_url": "https://preuve.ai/dl/codex-mcp-preuve-ai.mp4"
      },
      "publication": {
        "release_notes": "First release of this plugin. It replaces the earlier 1.0.2 submission, as the review team asked, with the MCP server connected during setup. Changes since that submission: the enrich_analysis tool description now describes each optional section in plain terms, and two promotional sentences were removed from the listing description."
      },
      "apps": "./.app.json"
    }
  }
}