I'm Krister, the founder of GQLens. I've spent the last four years shipping GraphQL APIs at companies of every shape: small startups where one service handled everything, and large orgs with dozens of federated subgraphs where no single human knew the whole schema. By day I'm a Technical Lead at a large org running on a federated GraphQL stack; outside of work I'm deep in LLMs, RAG systems, and the patterns that make AI agents actually useful on real codebases.
GQLens is the tool I wanted every time an AI agent tried to write a query against an API it had never seen. Instead of hallucinating fields, the agent asks GQLens, which knows the schema, the types, and the relations, and answers in a way the agent can turn directly into a valid query.
I write about the practical side of GraphQL at scale, MCP as a developer-tooling protocol, and the patterns that decide whether an AI-assisted codebase stays fast or collapses under its own context window.