
Bodhi
Bodhi: Secure, tokenized AI interactions on the Blockchain
Project links
About this project
Why are you participating for Based India?
Base is a new fast and efficient L2 which we targeted as our core blockchain in our project. Our application has a high throughput of transactions as multiple users are generating responses from LLMs concurrently and Base helps us to make that possible with very cheap gas and fast finality as it provides the security of Ethereum with scalability.
What challenges are you focusing on?
We are trying to make using LLMs cheap and affordable for everyone to use out there. Be it for large data analysis, or some day-to-day queries. You can select from various LLMs that are designed for each specific task and you only pay as much as you use that model. No need to subscribe monthly to 10 different models for dealing with different work.
How does your submission address this challenge?
Our platform solves this problem by allowing users to use any model of their choice depending on their use cases and pay as they use it. We provide a suite of LLMs and a great user interface to know how much a certain is prompt going to cost the user, a dashboard to show the token usage analytics, and much more. The best part about this is that the user pays with crypto and not fiat which makes it more accessible to everyone. We have our own ERC-20 token called “BOD” which users can buy with their stablecoin like USDC. We use Base as our layer 2 which makes our transactions fast and efficient while having the security of a blockchain like Ethereum.
Challenges we ran into
Problem 1: Obtaining real USDC on a test net for development can be difficult. Solution: Deploy your own ERC-20 token that mimics USDC's behavior for testing purposes. Problem 2: Solidity's lack of native floating-point numbers can make financial calculations challenging. Solution: Use fixed-point arithmetic, decimal libraries, or rational numbers to represent decimals accurately.
About the founders
Building on Base from India