Quasar
Agentic Trading on Base using Natural Language
Created on 21st February 2026
•
Quasar
Agentic Trading on Base using Natural Language
The problem Quasar solves
Executing token swaps on DEXs and conducting perpetual futures trades can be complex. It requires using intricate interfaces, which might be inconvenient for retail traders. With Quasar, we aim to allow autonomous perps trades, swaps, and prediction market trades using natural language. Our DeFi agent assistant will also be able to provide feedback on trades to prevent risky/poor trades. Quasar will allow for easier, quicker, and more intuitive trading on base, streamlining the process and bringing in a new wave of retail users currently sidelined by complex interfaces.
Challenges we ran into
We used GPT4o to parse natural language and turn it into instructions for the protocol. One problem we ran into with this was the LLM's inconsistency due to its randomized outputs; we had a few issues with the natural language that was supposed to result in hyperliquid trades. Additionally, we were planning on using Avantis, the perps DEX on base, for our perps trades. However, this was a problem for us because there was a 100 USDC minimum for trades on Avantis, which seemed a bit unrealistic for the retail traders that we are targeting. Thus, we pivoted to using Hyperliquid.
Use of AI tools and agents
We used GPT4o to parse the natural language.
Tracks Applied (5)
Futurllama
Best Use of Quicknode Hyperliquid HyperCore Streams
QuickNode
Best DeFAI Application
0g Labs
Base Self-Sustaining Autonomous Agents
Base
Integrate the Uniswap API in your platform
Uniswap Foundation
Technologies used