Find the bottleneck. Own the scarcity. Build the thesis.
Scarcity Scout is a structured investment research workbench for identifying, scoring, and stress-testing scarcity-driven investment theses — the kind of asymmetric setups where supply constraints meet surging demand and the market hasn't caught up yet.
Most investment tools help you screen stocks. Scarcity Scout helps you think.
Instead of starting with tickers, you start with a structural thesis — a bottleneck in the real world (energy permitting backlogs, semiconductor fab lead times, rare-earth processing capacity) — and work through a disciplined framework to decide whether it's investable, how to express it, and what would break the thesis.
The core workflow:
- Thesis → Define the structural shift, the bottleneck, and why now
- Evidence → Verify scarcity signals with source-backed evidence items
- Value Capture → Map the value chain to find who actually captures pricing power
- Expression → Build a layered portfolio from core holdings to speculative satellites
- Monitor → Set up confirming and disconfirming signals to track over time
Each step is guided by a synthesis block that tells you what the data says before you dive into the detail — so you spend time thinking, not scrolling.
One-click auto-fill generates a complete bottleneck analysis from a theme. Ollama is the default AI provider — run entirely against local open-source models with zero cloud dependency and full privacy. After every fill, an AI Analyst Memo surfaces the reasoning, fragile assumptions, and potential false friends — so the AI feels accountable, not magical.
AI analyses are automatically enriched with real-time context from 16+ financial news and commodity data feeds spanning energy, metals, mining, semiconductors, agriculture, and nuclear sectors (Reuters, CNBC, EIA, Mining.com, World Nuclear News, Investing News, SemiAnalysis, and more). Before the AI generates an analysis, the system fetches recent headlines relevant to your theme and injects them into the prompt — so the model reasons over current market conditions, not just its training data. Relevant headlines are also automatically added as formal evidence items in the scarcity evidence section with source attribution and confidence scores. Works with both local Ollama and cloud providers.
Add your own RSS feed URLs in Settings → Custom Data Feeds to tailor the research context to your specific domains. Custom feeds are merged with the built-in sources during every auto-fill, so the AI always has access to the latest data from the sources you trust most.
Prefer cloud models? Add your own OpenAI or Anthropic API key in Settings and specify any model your key supports. No built-in credits — you control your own usage and costs.
Nine-factor scoring system (scarcity severity, supply response speed, pricing power, barriers to entry, and more) with hover-over rubrics explaining exactly what a 1, 3, or 5 means for each factor. No ambiguous numbers.
Visual scatter plot of all your analyses on a scarcity-duration vs. market-mispricing grid. Instantly see which themes sit in the high-conviction quadrant.
A soft-gated portfolio builder that warns you if you're constructing positions before the thesis has earned it — minimum evidence thresholds, thesis breakers reviewed, confidence above baseline.
Structured checklist of what would kill the thesis (substitution risk, capital flood, demand decline, regulatory change) plus free-form disconfirming signals.
Map entities across six layers — demand creators, bottleneck owners, infrastructure, picks & shovels, operators, and integrators — to find where value actually accrues.
Full round-trip import/export for portability, version control, and team sharing.
Conversational AI assistant with full context of your analyses for ad-hoc questions, brainstorming, and deep dives. Works with Ollama or your own cloud API key.
Scarcity Scout is designed to be fully operable by AI agents — no UI required.
| Capability | Endpoint | Description |
|---|---|---|
| MCP Server | /functions/v1/mcp |
Native Model Context Protocol endpoint with tools for listing, creating, updating, deleting, and auto-filling analyses. Connect from Claude Desktop, Cursor, or any MCP client. |
| REST API | /functions/v1/api |
Standard CRUD with GET, POST, PATCH, DELETE and JSON responses. |
| OpenAPI Spec | /functions/v1/openapi |
Machine-readable API description for auto-discovery of endpoints, schemas, and operations. |
| Batch Operations | POST /functions/v1/api/batch |
Up to 50 create/update/delete operations in a single call. |
| Webhooks | /functions/v1/webhooks |
Subscribe to analysis.created, analysis.updated, analysis.deleted events with conditional filters (e.g., "confidence < 50%") and optional HMAC signature verification. |
| Research Context | /functions/v1/research-context |
Fetches recent news and data feed headlines relevant to a theme from 6+ financial RSS sources. Used automatically during auto-fill to ground AI in current market data. |
npm install
npm run devScarcity Scout defaults to a local Ollama instance for AI features. Install Ollama, pull a model, and start the server with CORS enabled:
# Install Ollama: https://ollama.com/download
ollama pull llama3.2
OLLAMA_ORIGINS="*" ollama serveThen open the app, go to Settings, click Test to verify the connection, and you're ready to auto-fill analyses.
If you prefer cloud models, go to Settings → AI Provider, select OpenAI or Anthropic, and paste your API key. No other setup needed.
Scarcity Scout uses Supabase for data persistence. To self-host:
- Create a Supabase project
- Run the migrations in
supabase/migrations/against your database - Set environment variables:
VITE_SUPABASE_URL— your Supabase project URLVITE_SUPABASE_PUBLISHABLE_KEY— your Supabase anon/public key
- Deploy the edge functions in
supabase/functions/to your project (only needed if using cloud AI providers via the chat or auto-fill features)
All AI calls are routed through your chosen provider — Ollama locally, or your own OpenAI/Anthropic key. There are no shared AI credits.
- Frontend: React · TypeScript · Vite · Tailwind CSS · shadcn/ui · Framer Motion
- Backend: Supabase (Postgres, Edge Functions)
- AI: Ollama (default, local) · OpenAI / Anthropic (BYO key)
- Thematic investors building conviction around structural supply/demand imbalances
- Analysts who want a repeatable framework instead of ad-hoc spreadsheets
- Portfolio managers stress-testing bottleneck theses before sizing positions
- Curious generalists who want to think about markets through a scarcity lens
MIT — Copyright (c) 2026 Drew Alan Hicks