The Graph Launches Amp: First Enterprise Blockchain Database

**PaySponge** has built a financial platform where every new agent automatically connects to **The Graph** without additional integration. - Agents like **Michaelgent** and **Clarabotagent** aren't standalone scripts—they're part of a unified infrastructure designed for the agent economy - Each new agent inherits the same routing path, eliminating setup friction - This represents a shift from hobbyist experimentation to platform-level growth The approach suggests a different scaling model: rather than individual agents building custom integrations, they plug into shared financial rails from day one.

The Graph has launched MCP (Model Context Protocol) servers that allow users to query live blockchain data using plain natural language—no GraphQL, SQL, or custom infrastructure required. **Key capabilities:** - Query Subgraphs and Substreams data conversationally - Works with Claude, Cursor, and Cline development tools - Eliminates technical barriers for non-developers - Enables real-time access to onchain data This builds on The Graph's existing infrastructure that has proven critical for AI agents requiring high-quality, real-time blockchain data. Previous implementations showed dramatic performance gains, with one analytics platform achieving 72,000% improvement over traditional RPC infrastructure. The natural language interface democratizes blockchain data access beyond developers to anyone who can ask a question, opening new possibilities for accounting, compliance, reporting, and rapid prototyping of data-driven applications. [Read the full technical breakdown](https://thegraph.com/blog/querying-blockchain-data-natural-language-mcp-skills/)

**AI agents operate at machine speed**, making thousands of decisions per minute in the crypto economy. Unlike humans who browse block explorers, these agents require **structured, normalized, and queryable blockchain data** to function effectively. - Agents assessing DeFi risk cannot work with raw transaction logs - They need data formatted for rapid querying and decision-making - The infrastructure enabling this agent economy operates behind the scenes This represents a fundamental shift in how blockchain data must be organized and accessed. As AI agents become more prevalent in web3, the demand for **properly structured data infrastructure** will intensify dramatically. The agent economy isn't about human-readable interfaces—it's about machine-optimized data pipelines that can support thousands of automated decisions every minute.

**Agent0 Subgraphs are now operational across five major blockchain networks** - Base, BNB Chain, Ethereum, Monad, and Polygon - using a single GraphQL schema. **Key features:** - Real-time indexing of every ERC-8004 agent across all chains - No code rewrites needed when switching between networks - Already processed over 1 million queries - Accessible via [Graph Explorer](https://thegraph.com/explorer?search=agent0) **What this enables:** - Developers can query agent data with a single GraphQL request instead of scanning thousands of blocks - AI agents can discover, hire, and interact with each other using structured blockchain data - Machine-scale data access for agents that operate continuously The infrastructure indexes agent identities, capabilities, reputation, and validation data in milliseconds. Developers building ERC-8004 agents can now access comprehensive onchain data without building custom indexers.

A new blockchain data tool called **graph-aave-mcp** now connects Claude AI with live Aave protocol data across seven chains. **Key capabilities:** - Simulate health factors during price crash scenarios - Track liquidation events in real time - Compare supply and borrow rates across different deployments - Monitor governance proposals through conversation The tool is powered by The Graph Protocol's indexed data and requires just one npm install to access every major Aave deployment. **Who benefits:** - Risk analysts can model scenarios conversationally - DeFi researchers gain real-time market insights - Builders can prototype dashboards by describing requirements - Content creators can query live data while writing The tool eliminates manual GraphQL queries, unit conversions, and stale data snapshots—users simply ask questions in plain English.