I Tried Making a One-Frame Comic Search Service Tiqav2 (Algolia + Cloudinary + Google Cloud Vision API)

Conversations with images are fun

Do you use emojis or images when communicating with chat tools? I use them often. When communicating with people, it feels stiff if it's just text. For example,

"OK, please proceed with that."

If it's just this phrase, it's hard to understand what kind of emotion the other person has.

Therefore, if you communicate with the following image, you can give a soft impression.

Tiqav2

What is Tiqav?

There is a service called Tiqav for having conversations using images.

http://dev.tiqav.com/

Currently, the service has ended. Tiqav2 was created with reference to that Tiqav.

What is Tiqav2?

Tiqav2 has two main features.

Save images and text
Search & display images

Two features

Save images and text

To search, we use the full-text search service Algolia.

https://www.algolia.com/

The information to be saved in Algolia is mainly three things: image URL and extension, and text. Images are saved in the image conversion & management service Cloudinary. After saving, you can get the image URL and extension from Cloudinary.

https://cloudinary.com/

Text is extracted by passing the image to the Google Cloud Vision API. Of course, you can also manually set the text.

https://cloud.google.com/vision/

Search & display images

Perform a full-text search with text. By combining the result ID and Extension, you can display the image. The types of Extensions are all those supported by Cloudinary.

"gif", "png", "jpg", "bmp", "ico", "pdf", "tiff", "eps", "jpc", "jp2", "psd", "webp", "zip", "svg", "mp4", "webm", "wdp", "hpx", "djvu", "ai", "flif", "bpg", "miff", "tga", "heic"

https://cloudinary.com/documentation/image_transformations#supported_image_formats

By using this image display feature, you can send images on Slack as follows.

For detailed features, please check the following repository.

https://github.com/silverbirder/tiqav2

SaaS is Ideal for Individual Development

This time, I left everything to SaaS for full-text search and image management. Text extraction was not necessary, but I used Google Cloud Vision API because it was quite usable.

When developing individually, resources (time, money, people) are very small compared to organizations. SaaS does one thing well, and many are free within the range of individual use. Unless you have a niche customization request, SaaS will meet most of your needs. If you want to know what kind of SaaS is available, please take a look at the site below. It should be helpful.

https://saasblocks.io/

Leaving the hassle to SaaS and focusing on product code is very important to me. By the way, the product code this time was made with CleanArchitecture + InversifyJS.