Android Police writer uses Google Gemini to catalog 400-game collection
By leveraging multimodal AI to convert photos into spreadsheets, a hobbyist bypassed niche subscription apps to build a voice-queryable database.
Akshay Bhalla, a writer for Android Police, recently used Google Gemini to automate the cataloging of a physical game collection exceeding 400 titles. The project demonstrates how multimodal AI can replace specialized software for inventory management.
To build the database, Bhalla took photos of game spines and covers in batches of 20 to 40. He then prompted Gemini to extract the visual data and convert it into Markdown tables, which were subsequently exported to Google Sheets. To ensure high quality and avoid the rate limits found in the free version of the AI, Bhalla utilized a Google AI Plus subscription. Once the data was housed in Sheets, he was able to query his entire collection using Gemini voice prompts.
The shift away from niche SaaS
Physical collectors frequently struggle to track large inventories that span multiple gaming platforms. While dedicated cataloging applications exist to solve this problem, many have transitioned to subscription-based models. For example, the iOS app Cibby limits its free tier to up to 10 games, prompting users to seek alternatives that do not require recurring payments for basic database functionality.
Implications for hobbyist data management
This workflow highlights a practical shift in how users interact with niche Software-as-a-Service (SaaS) products. By integrating Large Language Models (LLMs) with standard productivity tools like Google Sheets, users can create custom, searchable databases without the need for manual data entry or specialized software. "I avoided an app's expensive subscription model by asking Gemini to replicate the same feature," Bhalla stated.
Efficiency gains and future outlook
The time savings associated with AI-driven extraction are significant. Bhalla noted that the process took approximately one hour, whereas manual entry would have taken days. As multimodal AI continues to improve in its ability to parse text from images, similar workflows are likely to be adopted by collectors of books, vinyl, and other physical media to bypass restrictive app ecosystems.