Real-Time Stock Buyback Price Scraper
Budget: £20 – £250 GBP
Overview:
I need a custom-built data scraper that fetches live phone/tablet buyback prices from up to 30 UK-based websites, processes bulk Excel uploads, integrates with Custom GPT (Jarvis), and stores historical price data for long-term analytics.
The system must be:
• Lightweight
• Cost-free to run (use free-tier hosting or APIs)
• Fully automated
• 100% owned and controlled by me
⸻
What the System Must Do:
1. Scrape Live Buyback Prices
• From up to 30 UK phone buyback websites
• For Working and Faulty conditions
• Fast scraping (<10 sec/device ideally)
• Rotating proxies to avoid IP blocks
• Organized by manufacturer sections for performance
2. Expose a Free API/Webhook
• Accepts device name + condition
• Returns clean structured JSON:
3. Support Excel Uploads (up to 500 models)
• Handle batch pricing across 30 sites
• Return exportable Excel-style price comparisons
• GPT should handle these in chunks internally
4. Integrate with Custom GPT (Jarvis)
• Compatible with OpenAI’s function calling
• GPT should be able to:
• Fetch single device prices
• Handle multiple requests at once
• Summarise Excel uploads
• Recommend pricing actions
5. Log Historical Price Data
• Save every scrape with timestamp
• Store in free database (CSV, Firebase, or other free-tier)
• Structured for later trend analysis
6. Must Be Zero Ongoing Cost
• Host on Render, Railway, or Fly.io (free-tier)
• No Zapier, Make, or paid APIs
• Must provide full source code + handover docs
⸻
Deliverables:
• Fully working API + scraper hosted on free-tier platform
• Function-ready structure for GPT integration
• Excel-ready output format
• Lightweight logic to chunk large Excel files
• Logging of all scrapes into a historical database
• 100% ownership of all code
• Clear handover documentation or Loom video
⸻
To Apply:
Please include:
• Sample scraping projects (especially with IP rotation)
• What stack you’ll use (Python w/ Playwright preferred)
• How you’ll structure chunking & historical tracking
• Your time estimate for MVP (with 5–10 sites)
I need a custom-built data scraper that fetches live phone/tablet buyback prices from up to 30 UK-based websites, processes bulk Excel uploads, integrates with Custom GPT (Jarvis), and stores historical price data for long-term analytics.
The system must be:
• Lightweight
• Cost-free to run (use free-tier hosting or APIs)
• Fully automated
• 100% owned and controlled by me
⸻
What the System Must Do:
1. Scrape Live Buyback Prices
• From up to 30 UK phone buyback websites
• For Working and Faulty conditions
• Fast scraping (<10 sec/device ideally)
• Rotating proxies to avoid IP blocks
• Organized by manufacturer sections for performance
2. Expose a Free API/Webhook
• Accepts device name + condition
• Returns clean structured JSON:
3. Support Excel Uploads (up to 500 models)
• Handle batch pricing across 30 sites
• Return exportable Excel-style price comparisons
• GPT should handle these in chunks internally
4. Integrate with Custom GPT (Jarvis)
• Compatible with OpenAI’s function calling
• GPT should be able to:
• Fetch single device prices
• Handle multiple requests at once
• Summarise Excel uploads
• Recommend pricing actions
5. Log Historical Price Data
• Save every scrape with timestamp
• Store in free database (CSV, Firebase, or other free-tier)
• Structured for later trend analysis
6. Must Be Zero Ongoing Cost
• Host on Render, Railway, or Fly.io (free-tier)
• No Zapier, Make, or paid APIs
• Must provide full source code + handover docs
⸻
Deliverables:
• Fully working API + scraper hosted on free-tier platform
• Function-ready structure for GPT integration
• Excel-ready output format
• Lightweight logic to chunk large Excel files
• Logging of all scrapes into a historical database
• 100% ownership of all code
• Clear handover documentation or Loom video
⸻
To Apply:
Please include:
• Sample scraping projects (especially with IP rotation)
• What stack you’ll use (Python w/ Playwright preferred)
• How you’ll structure chunking & historical tracking
• Your time estimate for MVP (with 5–10 sites)