Case Study
Mobile app with AI: Scan Me
Case Study
Mobile app with AI: Scan Me
We developed a unique app using React Native and Google Cloud AI that lets users take a photo, ask a question, and receive a humorous, context-aware response. Over four months, we iterated through design and technical stages, including an early prototype of ChatGPT-vision, overcoming challenges with React Native's camera support. The result is a polished app with engaging visuals and analytics, now live on Google Play.
Project details:
Technologies: React native, Node.JS, React, MongoDB, Flask, Google Cloud AI
Project Goal: Our client came up with the idea of taking a photo of a person, asking a question, and receiving a humorous response, with the app also trying to infer characteristics from the photo and include them in the answer. We chose React Native for the technology stack, with Flask and MongoDB for the backend. The development was divided into iterations, with each iteration consisting of weekly sprints. We focused heavily on improving AI responses (prompt engineering) and enhancing the UX & UI. The total development time, including all iterations and updates, was 4 months.
Stage 1 - Improving idea planning and finding good references
The idea behind this app was very unique, so we conducted extensive competitor research on successful apps like Prisma to create an intuitive design. Additionally, we sought out good references for scanning animations to create a magical impression on users, making them believe that the algorithm is smart enough to understand secret questions just by analyzing photos.
Stage 2 - Technical demo, inventing our own ChatGPT-vision before its official release.
The next stage was to prove the technical feasibility of the idea. After conducting research, we arrived at the following solution:
Take a photo: Capture the image using the app.
Analyze it: Use Google Cloud AI to extract the context from the image.
Send text to ChatGPT-3.5: Formulate a custom prompt that includes the extracted image context and a question, then send it to ChatGPT-3.5
Get result: Receive and display the result from ChatGPT-3.5
This concept essentially anticipated the functionality of ChatGPT-vision, which was officially released two months later.
Stage 3 - Develop Mobile app using React Native.
We experimented with different design versions, focusing on improving the user experience. Additionally, we faced challenges with native camera support in React Native, which was quite a headache, but we managed to solve it!
Stage 4 - Mobile app functionality improvement and getting ready for the test!
The final stage involved polishing the app's visual aspects, integrating analytics (Google Firebase) with remote configs, creating creative advertisement videos, and finally uploading it to Google Play!
Project details:
Technologies: React native, Node.JS, React, MongoDB, Flask, Google Cloud AI
Project Goal: Our client came up with the idea of taking a photo of a person, asking a question, and receiving a humorous response, with the app also trying to infer characteristics from the photo and include them in the answer. We chose React Native for the technology stack, with Flask and MongoDB for the backend. The development was divided into iterations, with each iteration consisting of weekly sprints. We focused heavily on improving AI responses (prompt engineering) and enhancing the UX & UI. The total development time, including all iterations and updates, was 4 months.
Stage 1 - Improving idea planning and finding good references
The idea behind this app was very unique, so we conducted extensive competitor research on successful apps like Prisma to create an intuitive design. Additionally, we sought out good references for scanning animations to create a magical impression on users, making them believe that the algorithm is smart enough to understand secret questions just by analyzing photos.
Stage 2 - Technical demo, inventing our own ChatGPT-vision before its official release.
The next stage was to prove the technical feasibility of the idea. After conducting research, we arrived at the following solution:
Take a photo: Capture the image using the app.
Analyze it: Use Google Cloud AI to extract the context from the image.
Send text to ChatGPT-3.5: Formulate a custom prompt that includes the extracted image context and a question, then send it to ChatGPT-3.5
Get result: Receive and display the result from ChatGPT-3.5
This concept essentially anticipated the functionality of ChatGPT-vision, which was officially released two months later.
Stage 3 - Develop Mobile app using React Native.
We experimented with different design versions, focusing on improving the user experience. Additionally, we faced challenges with native camera support in React Native, which was quite a headache, but we managed to solve it!
Stage 4 - Mobile app functionality improvement and getting ready for the test!
The final stage involved polishing the app's visual aspects, integrating analytics (Google Firebase) with remote configs, creating creative advertisement videos, and finally uploading it to Google Play!
We developed a unique app using React Native and Google Cloud AI that lets users take a photo, ask a question, and receive a humorous, context-aware response. Over four months, we iterated through design and technical stages, including an early prototype of ChatGPT-vision, overcoming challenges with React Native's camera support. The result is a polished app with engaging visuals and analytics, now live on Google Play.
Project details:
Technologies: React native, Node.JS, React, MongoDB, Flask, Google Cloud AI
Project Goal: Our client came up with the idea of taking a photo of a person, asking a question, and receiving a humorous response, with the app also trying to infer characteristics from the photo and include them in the answer. We chose React Native for the technology stack, with Flask and MongoDB for the backend. The development was divided into iterations, with each iteration consisting of weekly sprints. We focused heavily on improving AI responses (prompt engineering) and enhancing the UX & UI. The total development time, including all iterations and updates, was 4 months.
Stage 1 - Improving idea planning and finding good references
The idea behind this app was very unique, so we conducted extensive competitor research on successful apps like Prisma to create an intuitive design. Additionally, we sought out good references for scanning animations to create a magical impression on users, making them believe that the algorithm is smart enough to understand secret questions just by analyzing photos.
Stage 2 - Technical demo, inventing our own ChatGPT-vision before its official release.
The next stage was to prove the technical feasibility of the idea. After conducting research, we arrived at the following solution:
Take a photo: Capture the image using the app.
Analyze it: Use Google Cloud AI to extract the context from the image.
Send text to ChatGPT-3.5: Formulate a custom prompt that includes the extracted image context and a question, then send it to ChatGPT-3.5
Get result: Receive and display the result from ChatGPT-3.5
This concept essentially anticipated the functionality of ChatGPT-vision, which was officially released two months later.
Stage 3 - Develop Mobile app using React Native.
We experimented with different design versions, focusing on improving the user experience. Additionally, we faced challenges with native camera support in React Native, which was quite a headache, but we managed to solve it!
Stage 4 - Mobile app functionality improvement and getting ready for the test!
The final stage involved polishing the app's visual aspects, integrating analytics (Google Firebase) with remote configs, creating creative advertisement videos, and finally uploading it to Google Play!
Other Projects
Other Case Studies
Check our other project case studies with detailed explanations
Other Projects
Other Case Studies
Check our other project case studies with detailed explanations
Other Projects
Other Case Studies
Check our other project case studies with detailed explanations