Changing the Game for Trusted Documents — What If Your Whole Platform Natively Supported It?
summary
Platform demonstrates a concrete workflow: GraphQL Codegen creates a manifest mapping operation hashes to documents, those manifests publish to Hive console, Hive Gateway pulls them via CDN and rejects any non‑registered queries, and CI/CD links app version to allowed operations so breaking changes are flagged. The team ships an MCP plugin that exposes annotated persisted documents as agent tools and connects to Claude to run queries and mutations, showing the pipeline works end‑to‑end in a pre‑recorded demo.
AI-written summary · about
chapters (22)
- 0:00Introduction to trusted documents
- 1:24GraphQL attack surface problem
- 2:29Trusted document workflow
- 4:16External trusted document storage
- 4:58GraphQL platform features
- 6:41Hive as trusted document store
- 7:04Trusted documents as MCP tools
- 8:08Demo setup overview
- 8:29Creating a Hive project
- 9:13Publishing subgraphs
- 9:57Hive gateway configuration
- 10:40Defining trusted operations
- 11:21App deployment process
- 12:06Verifying trusted documents
- 12:47Gateway trusted document setup
- 13:31Executing trusted operations
- 14:16MCP tool annotations
- 14:57MCP tool deployment
- 15:17MCP plugin installation
- 16:20Configuring MCP server
- 17:01Using MCP tools in agents
- 17:43Q&A session
related talks
01h 05mWorkshop: Scaling and Securing API Development with a GraphQL Platform
37mTransforming GraphQL: Full Stack Infrastructure as Code
35mGraphQL Docs: Beyond the Schema
35mSovereign docs
27mTransforming GraphQL: Infrastructure as Code for Front End Developers
29mClosing the Loop: How GraphQL Gives Coding Agents Eyes on What Actually Matters
18mLet Devs Be Devs Without Sacrificing Security
32mSponsored Session: Championing GraphQL Client in a Modern Platform