--- date: 2026-11-15 slug: graphlink-dart-huge-schemas categories: - Dart - Flutter description: "Set autoGenerateQueries to false and GraphLink generates only the operations you declare. Useful for huge schemas like GitHub or Shopify." id: D-11 demo_repo_path: "dart/d-01-quickstart" docs_page: "" --- # GraphQL Codegen for Huge Schemas in Dart: Turn Off Auto-Generated Queries By default, GraphLink reads your schema and generates a method for every query, mutation and subscription in it. For a bookstore that is exactly what you want. For a schema like GitHub's, GitLab's or Shopify's, it means a lot of generated code for operations your app will never call. This post shows the option that changes that: `autoGenerateQueries`. The video shows all of it, and the sections below follow the same order.

Watch on YouTube

## The default With the bookstore schema from the quickstart, the generated client has a method for each operation in the schema, for example fetching a book by ID and fetching the list of books. You never had to write a query for them. That is the default: `autoGenerateQueries` is `true`. ## Turn it off In `config.json`, under the Dart client config, set it to `false`: ```json { "schemaPaths": ["schema.graphql", "queries.graphql"], "mode": "client", "outputDir": "lib/generated", "clientConfig": { "dart": { "packageName": "bookstore_quickstart", "autoGenerateQueries": false } } } ``` In the video I delete the `lib/generated` folder first, so nothing old is left, and run the generator again: ```bash glink -c config.json ``` ## What is left Before, the client had the book-by-ID and books methods. Now the generated queries class only has what you declared yourself, in the `queries.graphql` file from the previous video (`getBooksWithoutAuthors`). Everything else is gone. Two things to know from the video: - The file does not have to be called `queries.graphql`. It can be any file with any extension, as long as it is listed in the config next to the schema. - Set the option back to `true`, or remove it, run the generator again, and the other methods come back. Nothing is lost, it is only a choice about what to generate. ## Which one should you use My rule of thumb, from the video: - **Small schema:** keep the default. You don't have to declare every query, and in general you want to be able to get every field. - **Huge schema:** turn it off and declare only the operations your app uses. The GitHub, GitLab and Shopify schemas are in the `examples` folder of the GraphLink repo, if you want to try it on something big. ## What this video does not cover I did not run a real GitHub schema in the video, and I did not cover the other options for huge schemas. This one is the on/off switch. ## Try it - Code from the video: [graphlink-demos](https://github.com/Oualitsen/graphlink-demos) - The project: [github.com/Oualitsen/graphlink](https://github.com/Oualitsen/graphlink). A star on GitHub is the easiest way to help. - Docs: [graphlink.dev](https://graphlink.dev)