Are you currently operating a Subgraph and finding yourself constrained by its boundaries? There is a convenient solution to upgrade your workflow. The substreams-convert utility automatically translates your schema.graphql entities into protobuf types. Additionally, it shifts your AssemblyScript handlers into map and store modules while ensuring your initialBlock is perfectly preserved throughout the transition. The end result gives you the exact same data, transformed into a flexible stream that you can direct to any sink you need.
By utilizing the Subgraph MCP, your AI agents can access and explore onchain data at a significantly faster pace. Transitioning an agent from a basic protocol name to a live query on the Subgraph MCP requires a straightforward sequence of exactly four calls. The process begins by searching for subgraphs via a keyword. Next, the agent retrieves the deployment 30D query counts, followed by obtaining the schema, and concludes by executing the query. Interestingly, evaluating the query volume is a crucial phase that many agents tend to bypass. Including this specific volume check is the exact measure needed to ensure they avoid interacting with dead deployments.
The prediction market industry is seeing explosive growth, having reached $150B+ in cumulative volume. Behind the scenes, every trading agent, dashboard, and market operating in this space depends completely on indexed onchain data. In the case of Polymarket, all of this essential information is already indexed by The Graph. You have multiple options to interact with it, whether you choose to make queries using Subgraphs, stream the data through Substreams, or simply call the Token API.
Read further details here: https://thegraph.com/blog/prediction-markets-data-layer/
The AI coding assistant has been upgraded in a recent weekend build and now comes equipped with ten specific Substreams skills. By simply detailing the on-chain data you require, you can prompt the tool to automatically scaffold the module, wire the sink, and initiate the run. Because of this streamlined workflow, developers are able to retrieve live blockchain data almost as quickly as they can draft the initial project specifications.
In order to successfully operate on-chain, an AI agent relies on three essential capabilities: awareness, identity, and payment. The Graph provides the underlying infrastructure for two of these core requirements. By leveraging Subgraphs and MCP, raw and previously unsearchable blocks are transformed into clear context that an agent can easily process. Additionally, x402 enables the agent to pay per query directly, entirely removing the need for an API key. Together, these technologies form the foundational stack beneath the agentic web.
We are conducting a quick survey for developers who create agents. What is the most frequent cause behind your system breaking down?
Are you dealing with flawed or missing data, a lack of payment methods to acquire data, or complications involving identity and permissions? Please respond to share which of these obstacles causes the greatest frustration.
今年初めに、The Graph の将来について語る2部構成シリーズの第1部として「テクニカル・ロードマップ」を The Graph Foundation が公開しました。いま、パート2をお届けできることを大変うれしく思います。本更新では、新たな Foundation のミッション(委任)に関する正式な宣言と、そのロードマップを Foundation がどのように実行する意図かについての明確な説明をお伝えします。以下のリンクから、発表の全文をご覧いただけます。