AI Shovel Radar

AI Shovel Radar finds monetizable opportunity signals in the AI gold rush.

It is built for people who want to make money with AI but do not want another vague list of "10 side hustles." The Actor scans public conversations, extracts recurring pain and ambition, then turns that into practical shovel ideas: services, APIs, automation offers, template packs, niche reports, and small tools people can test.

It does not promise income. It shows you where to dig.

This is a separate product from HN Opportunity Finder. HN Opportunity Finder is for developer pain and technical leads. AI Shovel Radar is for ambitious AI-money seekers who may not know or care what Hacker News is; Hacker News is only one optional signal source under the hood.

What this Actor gives you

Each dataset row is a shovel card:

• who the target buyer might be
• what they appear to want
• what pain or confusion is visible
• what shovel you could sell into that demand
• a concrete first offer
• tools needed
• evidence from public conversation
• lead quality: research, warm_lead, or direct_lead
• beginner difficulty
• saturation risk
• failure modes
• next step

Best first run

{
  "ambitionLevel": "side_hustle",
  "preferredModels": ["service", "api_actor", "template_pack"],
  "skillLevel": "beginner",
  "budget": "zero_to_low",
  "timeAvailable": "weekends",
  "interests": ["automation", "local business", "content"],
  "dataSources": ["apify_store", "reddit"],
  "redditCommunities": ["sidehustle", "Entrepreneur", "smallbusiness", "ChatGPT", "SaaS"],
  "days": 30,
  "maxResults": 100,
  "outputLimit": 10,
  "minScore": 45,
  "leadQualityMode": "all",
  "includeHigherDifficulty": false
}

Output example

{
  "opportunityTitle": "AI workflow automation around support, chatbot",
  "goldRushCategory": "AI customer support automation",
  "monetizationModel": "service",
  "targetBuyer": "small teams with repeated customer questions and no dedicated support engineer",
  "buyerDesire": "They want faster replies without hiring support staff or learning AI tooling.",
  "visiblePain": "People describe friction around setup and support, which can be packaged into setup, workflow, or tooling help.",
  "shovelToSell": "A fixed-price AI support FAQ and triage setup.",
  "firstOffer": "I'll turn your existing help docs and inbox patterns into a basic AI support triage flow for $199.",
  "difficulty": "beginner",
  "timeToFirstAttempt": "1 weekend",
  "toolsNeeded": ["Crisp or Tidio", "OpenAI or Claude", "Google Sheets", "Zapier or Make"],
  "saturationLevel": "medium",
  "leadQuality": "warm_lead",
  "leadQualityReason": "Operational pain is visible, but the buying/request intent is not explicit enough.",
  "directLead": false,
  "scores": {
    "buyerPain": 72,
    "monetizationClarity": 75,
    "beginnerFeasibility": 70,
    "competitionRisk": 55,
    "evidenceStrength": 64,
    "overall": 67
  },
  "failureModes": [
    "Support bots fail if the source FAQ is thin or wrong.",
    "Buyers may worry that AI will annoy customers.",
    "The offer needs clear human handoff, not full autopilot hype."
  ],
  "nextStep": "Pick one business with public FAQs and build a small before/after demo from their actual questions."
}

Input tips

Use interests to steer the scan:

• automation
• local business
• content
• lead generation
• customer support
• data
• SEO
• video clipping
• RAG
• AI agents

Use data sources to choose where the signal comes from:

• apify_store: stable, free marketplace supply/demand signal from existing Actor categories, descriptions, usage, and pricing
• hackernews: optional developer/news-heavy fallback signal
• reddit: broader buyer-psychology signal from targeted public subreddit search; anonymous access can be sparse or rate-limited, so the Actor treats Reddit blocks as an empty source instead of crashing

Use redditCommunities to steer Reddit toward the actual AI-money audience. Good starter communities are sidehustle, Entrepreneur, smallbusiness, ChatGPT, and SaaS. This keeps the product focused on ambition/shovel demand rather than generic AI commentary. Add all only if you explicitly want broader Reddit search noise.

Use preferred models to control the kind of shovel:

• service: setup, audit, done-for-you workflow
• api_actor: APIs, Apify Actors, data tools
• template_pack: prompts, Notion systems, workflow packs
• no_code_automation: Zapier/Make style automations
• content_newsletter: attention and audience plays
• data_report: niche reports and lead lists

Use lead quality mode to control strictness:

• all: broad opportunity research and gold-rush signals
• qualified: only warm_lead and direct_lead
• direct_only: strict explicit buyer/request intent; this will usually be sparse on HN

directLeadOnly: true is a compatibility shortcut for leadQualityMode: direct_only.

The direct-lead classifier is intentionally conservative. High score often means “interesting AI-money discussion,” not “someone to contact.”

Current data sources

Version 0.4 scans Apify Store marketplace search and targeted Reddit communities by default. Hacker News remains available as an optional fallback source via dataSources. No API key, no proxy, and no paid data dependency are required.

Future versions can add Product Hunt, GitHub discussions, YouTube, X, and Indie Hackers-style signals.

Important disclaimer

This Actor does not provide financial advice and does not guarantee income. It outputs public-data opportunity signals and practical first-test ideas. What you do with those ideas is up to you.

Local API mode

cd ai-shovel-radar
python -m venv .venv
. .venv/bin/activate
pip install -e '.[test]'
uvicorn ai_shovel_radar.app:app --app-dir src --host 0.0.0.0 --port 8000

Example:

http://localhost:8000/v1/shovels?interests=automation,local%20business&dataSources=hackernews,reddit&redditCommunities=sidehustle,Entrepreneur,smallbusiness&preferredModels=service,api_actor&outputLimit=10