• 

  KawAPI server with transactions

KawAPI

From the Japanese 河 (Kawa), which translates to stream.

KawAPI is a bridge between data consumers and data producers. It is a tool that allows the monetization of any data stream - whether it be industry or citizen scientists.

Inspiration

We believe there are industries with large datasets that have non-trivial value, and no easy way to open up these datasets. We'd like to create a way to value and provide access to these datasets, with dynamic pricing based on the consumer and producer requirements.

An example: machine learning and neural networks are solving real world problems, and they require large quality real world datasets. In general, the larger the dataset you have to train a neural network, the better the neural network will perform. The same is true of data uniqueness, which many small players can contribute to.

Doctors, researchers, small business owners, even large industrial companies, all have incredibly valuable data. This data is locked behind closed doors, because they have no easy way to monetize their data.

Facebook and Google have made people realize that data is gold. We're providing a tool to mine that gold and open it up to the world.

When Google gives you a Google Form to request access to paid APIs, that's when you know the current system is broken: https://docs.google.com/forms/d/e/1FAIpQLScK9o4sb96aJMRsK9UUCAiKLjQRJGyrBg2d9XKSL9rOFfpOzg/viewform

Give it a try

Open http://www.kawapi.io/37c1e3be-da52-4f85-b97d-b407639ffe7b/endpoint/molecule/api/project in your browser. You'll get 10 free transactions per session before the server responds that you need to start sending IOTA micro-transactions.

The Tech

KawAPI uses IOTA to perform microtransactions for dynamically priced data streams. It consists of a client CLI that securely manages the transactions from your IOTA wallet, and a server that acts as a reverse proxy between the consumer and the producer, to securely provide access to datasets and APIs.

The server is hosted at kawapi.io, but could be self hosted. Your seed never leaves your side, and all code is open-source.

• rate limiting, pagination, free preview tiers, etc.
• value datasets by access bandwidth rather than selling the actual dataset
• if you have a machine learning model that needs access to the dataset at a high tx rate, you'd pay more than someone who wanted a sample of the first 100 entries
• pay for computation CPU load

Challenges

Both our team members were unfamiliar with Go and IOTA, and had to learn these during the Hackathon! It was a great challenge, and that's what Hackathons are all about... learning.

Roadmap

• Implement a Ngrok/Chisel style tunnel to allow easy provision of local datasets.
• Provide interfaces to decentralized datastores, like Ocean, OrbitDB, IPFS, and even IOTA itself.
• Create a marketplace of API endpoints, open for producers to add their endpoints, and consumers to discover endpoints.

Built With

• go
• iota
• ocean