GrantSage

GrantSage is a data-native AI agent that helps Web3 builders, researchers, and funders explore funding patterns, compare ecosystems, and uncover insights from grant datasets. Powered by a custom agent deployed on Autonome, GrantSage is connected to a curated set of JSON-LD grant metadata from Gitcoin, Optimism, and Octant via DAOIP-5. Users can ask GrantSage analytical questions such as:

"Which project received the highest total funding in Optimism?"

"How many projects got funding from more than two platforms?"

"Is there a correlation between grant size and approval count?"

It provides instant, context-aware responses based on structured data, giving ecosystem stakeholders clarity and insight into the flow of public goods funding.

GrantSage is built using:

Autonome + Eliza Framework for deploying the AI agent

OpenRouter as the model provider

JSON-LD grant data from DAOIP-5, hosted on GitHub and structured using the EIP-4824 schema

A React + Tailwind SPA for a clean user-facing interface

Basic Auth-protected API for secure communication with the deployed agent

We used CSV-to-JSON converters to transform raw data into linked metadata for effective semantic querying. The agent uses prompt-engineered examples to perform analytical reasoning on grant sizes, funding rounds, and project overlaps. The frontend pulls messages from the agent API and displays them in a smooth, minimal interface using your preferred color palette.

Bonus: the project is fully open source and modular — designed for extension to other datasets or protocols