SiliconSignal — Automated Semiconductor Tracking Tool

Real-time semiconductor supply chain intelligence for automated risk assessment and lifecycle monitoring.

Mission Critical Supply Chain Visibility

SiliconSignal is a high-precision monitoring platform designed to detect logistics risks, lead-time shifts, and lifecycle transitions in real-time. By leveraging the Tinyfish web agent (Mino), it extracts live signals directly from foundry bulletins and primary distributor channels.

---

System Interface

---

1. Technical Framework

The system operates as a distributed data collector, mapping part-level signatures to identified web sources.

Data Acquisition Strategy

Stage	Technical Operation	Purpose
Source Mapping	Heuristic identification of relevant foundry/distributor URLs.	Minimize scan latency and maximize signal relevance.
Web Tracking	Execution of headless browser instances for multi-step navigation.	Bypassing static caches to reach live inventory and status pages.
Signal Extraction	DOM-level parsing of unstructured lead times, stock levels, and MOQ.	Converting fragmented web data into structured technical metrics.
Logic Assessment	Rule-based comparison against historical snapshots.	Detecting factual deviations (e.g., NRND status change).

Output Data Schema

{
  "tracking_metrics": {
    "part_number": "STM32F407VGT6",
    "lifecycle": "NRND",
    "lead_time": 18,
    "availability": "Limited"
  },
  "logistics_risk": {
    "score": 75,
    "level": "HIGH",
    "reasoning": "Detected 4-week lead-time spike compared to baseline + NRND signal at source."
  },
  "telemetry_logs": [
    "[Tinyfish] identified 3 sources: DigiKey, Mouser, TI Direct",
    "[Tinyfish] Pricing: Detected price point around $5.20"
  ]
}

---

2. Integration & Usage

API Implementation

SiliconSignal exposes a robust REST API for integration into procurement and PLM workflows.

cURL Example

curl -X POST "http://localhost:3000/api/scan" \
  -H "Content-Type: application/json" \
  -d '{"part_number": "STM32F407"}'

TypeScript Implementation

const fetchRiskProfile = async (part: string) => {
  const res = await fetch("/api/scan", {
    method: "POST",
    headers: { "Content-Type": "application/json" },
    body: JSON.stringify({ part_number: part }),
  });
  return await res.json();
};

---

3. System Architecture

Data Flow Pipeline

graph TD
    User([User]) -->|Inputs Part #| DP[Dashboard / PlatformView]
    
    subgraph "Core Backend"
        API[API: /api/scan]
        Store[(Historical Snapshot Store)]
        Engine[Technical Assessment Engine]
    end

    subgraph "Tracking Layer (Tinyfish/Mino)"
        Crawler[Automated Crawler]
        DOM[DOM Extraction Engine]
        Sources((Live Web Sources))
    end

    DP -->|Request| API
    API -->|Deploy| Crawler
    Crawler -->|Navigate| Sources
    Sources -->|Telemetry| DOM
    DOM -->|Parsed Data| Engine
    Store <-->|History Link| Engine
    Engine -->|Structured Report| API
    API -->|Result| DP

Loading

Monitoring Workflow

sequenceDiagram
    participant U as Client UI
    participant S as Scan Orchestrator
    participant M as Tinyfish (Mino Web Agent)
    participant E as Assessment Engine

    U->>S: Track Part Request
    S->>M: Action: Scan Distribution Channels
    M-->>M: Navigate & Parse Stock/Price
    M->>S: Raw Scrape Response
    S->>M: Action: Scan Foundry Lifecycle
    M-->>M: Navigate bulletins & Alert Logs
    M->>S: Raw Lifecycle Data
    S->>E: Process Signals & History
    E->>U: Final Technical Report

Loading

---

Key Capabilities

• Live Web Verification: Real-time checking of foundry and distributor pages for direct status signals.
• Logbook Transparency: Dedicated terminal logs showing exact tracking steps and identification success.
• Logistics History: Persistence layer to track changes in lead times and status over months.
• Industrial Aesthetic: Premium dark-mode interface designed for professional engineering environments.

---

Engineering Standards

• Concurrency: All outbound requests use timeouts, retries, and capped parallelism.
• Input Validation: Part numbers are normalized and validated before scan execution.
• Caching: Recent scans are cached with TTL to reduce repeated work.
• Signal Priority: Explicit signals override inferred heuristics.
• Readability: Shared helpers and clear log messages for maintainability.

---

Scan results and user feedback

• Sample parts: The scan form includes one-click sample parts (e.g. NE555, ATmega328P, STM32F103C8T6) that typically return lifecycle and availability from distributor scans.
• No N/A in main fields: When a scan finds at least one source, lifecycle shows parsed value or “Active”; availability, price, and lead time show parsed values or “—” when not found.
• Traceability Evidence: Use the Ref links under each result to open distributor pages for price and lead time when those fields show “—”.
• Manufacturer: Filling the optional manufacturer field (e.g. Texas Instruments, Microchip) can improve parsing. The “lacks manufacturer information” message only appears when no distributor sources were found.

---

Getting Started

Environment Setup

Create a .env.local in the frontend directory:

MINO_API_KEY=your_key_here

The Tinyfish/Mino tracker runs without API keys, but adding MINO_API_KEY enables enhanced telemetry logging.

Running Locally

cd frontend
npm install
npm run dev

If port 3000 is already in use, stop the existing process or run with a different port. PowerShell example:

$env:PORT=3000; npm run dev

---