Engineering
October 16, 2023
3 min read

Kong Hackathon 2023 Winners Showcase Innovative Uses of AI for APIs

Kaitlyn Barnard

We’re thrilled to announce the winners of the 3rd Annual Kong Hackathon, held in conjunction with API Summit 2023! This year's hackathon centered around API innovation using artificial intelligence, challenging participants to create AI-powered solutions revolving around APIs.

Over the past month, we’ve been astounded by the creativity, innovation, and problem-solving demonstrated through the many excellent submissions from the Kong Community. After careful consideration, our judges have selected the top contributions that showed unique applications of AI to enhance and extend API capabilities.

Read on as we highlight our 2023 Kong Hackathon winners and their submissions:

1st Place: Alparslan Avci

This plugin enhances API responses by integrating with OpenAI's ChatGPT API. It intercepts API responses, uses ChatGPT to enhance the responses into more natural language, and returns the improved output to clients. 

In the demo, the plugin turns technical JSON errors into human readable form using ChatGPT prompts. The demo shows how the plugin harnesses AI to transform raw API responses for a better client experience.

2nd Place: br-chimarrao

The Kong internationalization plugin is a powerful tool that allows for real-time translation of API responses into over 7,000 languages. It achieves this by utilizing AI models such as DL Translate and LinguaPi. When a response is intercepted, the plugin sends it to a Python engine for translation and then returns the translated content to the client. 

In the demo, the plugin was used to translate Portuguese API responses into English, Arabic, and Italian in real-time. This was done by using language headers and supporting multiple payload fields. Additionally, the plugin offers text-to-speech capabilities for accessibility purposes. The use of specialized AI models further improves the performance of the plugin.

3rd Place: QuadAI

This submission aims to enhance the search functionality of Kong's Developer Portal using AI. The integration of OpenAI chat capabilities and Azure cognitive search will allow for the indexing and querying of API specifications. This will improve the discoverability of large and complex APIs. 

In the demo, the plugin is shown to handle the massive Spotify API by chunking it for cognitive search. This feature enables OpenAI to respond to queries on specific parts of the API specification in the Kong Developer Portal, making the process of discovery easier.

4th Place: Verifone GSC

The Custom Plugin project uses machine learning modules in a Kong plugin to allow natural language queries for AI-powered analytics of unstructured data, such as logs. This provides conversational insights from raw text. 

The demo showcases the plugin analyzing a Kong error log file to answer natural language queries about upstream health and duration of downtime. This demonstrates the plugin's conversational abilities to parse unformatted data and return human-readable insights.

5th Place: AI Plugin Guru

AI Plugin Guru is a Kong Developer Portal add-on that accelerates plugin development by automatically generating customized Kong plugin code using ChatGPT, reducing the process to prompt-based conversation.

The demo prompts ChatGPT to create plugins by retrieving the installed plugins and developer API key, handling the end-to-end workflow that enables easy plugin generation with just natural language conversation.

Thank you! 

In addition to the top 5 winners, we had many other fantastic submissions that were awarded digital contributor badges and helped earn donations to the Diane Fossey Gorilla Fund.

We thank everyone who participated in the 2023 Kong Hackathon on AI and API innovation. Our community continues to amaze us with your skills, ideas, and willingness to learn new technologies. We can't wait to see what you come up with next year!

Stay tuned for more updates from Kong and API Summit 2023. And congratulations to all of our 2023 hackathon winners!

Tags:AI