Hey Graph Community,
As The Graph grows with more subgraphs being deployed, it’s becoming harder for users—especially newcomers—to find the right subgraph for their data needs or determine if they need to build one. I’d like to propose a feature to improve discoverability and usability: Subgraph Finder, an AI-powered tool integrated into Graph Explorer.
I don’t have the skillset to implement this myself, but wanted to suggest it in case it sparks interest.
Here’s the idea:
The Problem
With thousands of subgraphs across networks, users face friction:
-
Finding a subgraph that indexes the contract or data they care about.
-
Understanding what data a subgraph provides without diving into its schema.
-
Deciding whether to use an existing subgraph or create a new one.
This slows onboarding and adoption, which we can fix with smarter tools.
The Solution: Subgraph Finder
A step-by-step tool in Graph Explorer to pinpoint subgraphs based on user needs:
- Filter by Network
-
Status: Already exists in Graph Explorer (e.g., Ethereum, Arbitrum).
-
Action: Users select their network of interest.
- Filter by Smart Contract
-
Status: Feasible with existing tools like Mike from DataNexus’s subgraph (FMTUN6d7sY2bLnAmNEPJTqiU3iuQht6ZXurpBh71wbWR ), which searches manifests by contract address. I’ve built a dashboard using it here: subgraph-search-by-contract.vercel.app.
-
Action: Users input a contract address (e.g., 0xd829c1d3649dbc3fd96d3d22500ef33a46daae46), and it returns matching subgraphs (e.g., Monadex).
- Retrieve Schemas
- Status: Possible via the Graph Network Subgraph. Here’s a sample query:
graphql
query {
subgraphDeployments(first: 5, where: { id_in: ["A3Np3RQbaBA6oKJgiwDJeo5T3zrYfGHPWFYayMwtNDum"] }) {
id
ipfsHash
manifest {
schema { schema }
manifest
}
versions { subgraph { id owner { id } } }
}
}
- Action: Fetches the schema.graphql contents (e.g., Monadex’s Pair, Token) for filtered subgraphs.
- AI-Assisted Matching
-
Status: New feature using an LLM (e.g., Grok or similar).
-
Action:
-
User inputs their need (e.g., “I want trading volume and liquidity for a DEX on Arbitrum”).
-
AI analyzes schemas from Step 3, matches them to the intent, and responds:
- Match: Suggests a subgraph (e.g., Monadex) with a sample query:
-
query {
pairs(first: 10, orderBy: volumeUSD) {
id
volumeUSD
reserveUSD
}
}
Or provide
* No Match: “No subgraph exists. Check [docs](https://thegraph.com/docs) to build one.”
Benefits
-
Ease of Use: Simplifies subgraph discovery for all skill levels.
-
Adoption: Lowers barriers, encouraging more dApp development.
-
Scalability: Handles growing subgraph counts with AI efficiency.
Why It’s Feasible
-
Builds on existing tools: Graph Explorer’s network filter, Mike’s subgraph, and the Graph Network Subgraph.
-
My dashboard proves Step 2 works.
-
The query in Step 3 already retrieves schemas (tested with Monadex).
-
AI integration is the main lift—leverage tools like xAI’s Grok for schema parsing and query generation.