project screenshot 1
project screenshot 2
project screenshot 3
project screenshot 4
project screenshot 5
project screenshot 6

Talk to your dao

DIAO (pronounced like “ciao”) is an AI-powered DAO assistant that enables intuitive interaction with your DAO while displaying data visually instead of just text.

Talk to your dao

Created At

Agentic Ethereum

Project Description

DIAO (pronounced like “ciao”) is an AI-powered DAO assistant that enables intuitive interaction with your DAO while displaying data visually instead of just text.

Key Features: 1. ENS Domain Integration • Lookup ENS domains and render them as visually appealing profile cards. 2. Subgraph Query Generation • Automatically generate queries based on the subgraph schema (currently supports only Nouns DAO). 3. Proposal Listing • Fetch and display a clickable list of DAO proposals for easy exploration. 4. Detailed Proposal View • Retrieve a specific proposal by ID, displaying it as: • A card-style summary UI. • A full document-style detailed view.

Additionally, DIAO integrates general LLM capabilities, enabling conversational interaction and deeper discussions about the displayed DAO data.

How it's Made

Templates: • Vercel’s AI SDK • Vercel’s AI Chatbot Template • Shadcn Components

Partners: • The Graph • ENS (likely not a partner this time) • Nouns (not a partner this time, but their subgraph provided the necessary data)

Hacky Solutions: I wasn’t initially planning to generate queries through the LLM. However, I became frustrated navigating The Graph’s subgraph explorer UI. So, I found the schema, fed it into an LLM, and asked it to create a query. Surprisingly, it worked well enough that I integrated it into the app as a tool.

I hardcoded the Nouns DAO subgraph schema because it was the one I was working with. If I had more time, I would’ve made it dynamic—capable of finding the schema for any subgraph and generating any query.

Next Steps (if I had more time, my fault for not starting til Wednesday) I would have multi-agent orchestration. Instead of simply returning the GraphQL query code, it would: 1. Take any user question. 2. Identify the relevant subgraph. 3. Fetch the schema. 4. Generate and execute the query. 5. Return a summary or formatted visualization of the data.

background image mobile

Join the mailing list

Get the latest news and updates