Compare grocery prices across stores from a list
A tech worker built a Gemini browser agent that takes a typed or scanned handwritten grocery list and compares prices on Amazon Fresh and Flipkart Minutes. Task: Suggested implementation: Build a browser automation agent using Playwright and Gemini 2.0 Flash to compare grocery prices. The agent accepts typed or scanned lists, searches Amazon Fresh and Flipkart Minutes, captures price screenshots, and selects the best option. Human role: Review the selected items and confirm before adding to cart, as the source does not specify automated checkout.; tools: Gemini. Human role: Source evidence is unavailable regarding specific human roles or manual checks required for this workflow.; limitations: The published code is simplified and lacks clear input handling for lists. Integration with real checkout flows is unproven. Source reliability is low due to a 403 error on the original URL and lack of verifiable metrics. Human approval is required for any purchase actions. Proposed advice: Test with a small, non-purchasing list first to validate scraping accuracy..
Tools
Published on Agent Work Atlas:
Overview
A tech worker built a Gemini browser agent that takes a typed or scanned handwritten grocery list and compares prices on Amazon Fresh and Flipkart Minutes.
Workflow
Suggested implementation: Build a browser automation agent using Playwright and Gemini 2.0 Flash to compare grocery prices. The agent accepts typed or scanned lists, searches Amazon Fresh and Flipkart Minutes, captures price screenshots, and selects the best option. Human role: Review the selected items and confirm before adding to cart, as the source does not specify automated checkout.
Reported result
The author claims the tool saves weekend time and finds good deals, but no specific metrics for time saved, cost reduction, or accuracy are provided. A reader noted issues, suggesting potential reliability gaps. Proposed acceptance check: Measure time spent on manual comparison vs. agent-assisted selection for 10 items, and verify price accuracy against live site data.
Human role
Source evidence is unavailable regarding specific human roles or manual checks required for this workflow.
Limitations
The published code is simplified and lacks clear input handling for lists. Integration with real checkout flows is unproven. Source reliability is low due to a 403 error on the original URL and lack of verifiable metrics. Human approval is required for any purchase actions. Proposed advice: Test with a small, non-purchasing list first to validate scraping accuracy.
Try it
Act as a browser automation engineer. I need to build a prototype for a grocery price comparison agent. Inputs: 1. A grocery list file [replaceable: e.g., 'grocery_list.csv' or image of handwritten list]. 2. Target platforms: Amazon Fresh and Flipkart Minutes. 3. Output location: [replaceable: e.g., 'results/comparison_report.json']. Steps: 1. Parse the input list into individual items. 2. Use Playwright to open a headless Chrome instance. 3. For each item, search on both platforms. 4. Extract price, discount, and delivery time data via DOM parsing or screenshots. 5. Compare options and select the best value. 6. Generate a JSON report with item, platform, price, and link. Safeguards: - Do NOT add items to cart or checkout. - Perform a dry run with 3 items first. - Ask me to confirm the parsed list before starting the search. - If a site blocks access, log the error and skip that item. Deliverables: - Python script using Playwright and Gemini API. - Sample output JSON. - Instructions for running the script.
Source reports are not independently verified. Publication date is the catalog publication date, not the original article date.