Ai

Axolotl CLI enables to use prepared AI prompt to generate resolvers code and (soon) generate test code for those resolvers

Generate graphql schema

Provide OPEN_AI_API_KEY env variable and run

npx OPENAI_AI_API_KEY=xxx axolotl gai  "schema for booking.com like system"

This command will generate the graphql schema

Config params

prompt_info - you can add 'baked' additional prompt information into your axolotl config file

{
    "graphql_prompt_info": "use pipes",
    "agent_model": "gpt-4.1-nano"
}

Prompt info can be also a txt file for bigger prompts

{
    "graphql_prompt_info": "graphql.txt",
    "agent_model": "gpt-4.1-nano"
}

Generate backend resolver

Given schema.graphql

type User{
    username: String!
    posts: [Post!]
}
type Post{
    title: String!
    content: String!
    user: User!
}

Provide OPEN_AI_API_KEY env variable and run

npx OPENAI_AI_API_KEY=xxx axolotl ai path/to/schema.graphql User posts "resolver to list all the posts user has created use mongodb"

This command will generate the axolotl compatible typescript backend code for the resolver and copy it to clipboard

Config params

prompt_info - you can add 'baked' additional prompt information into your axolotl config file

{
    "prompt_info": "use postgres with drizzle library to manage database",
    "agent_model": "gpt-4.1-nano"
}

Generate frontend react component

This agent generates react components that communicate with provided backend.

Given schema.graphql

type User{
    username: String!
    posts: [Post!]
}
type Post{
    title: String!
    content: String!
    user: User!
}
type Query{
    posts: [Post!]
}

Provide OPEN_AI_API_KEY env variable and run

npx OPENAI_AI_API_KEY=xxx axolotl fai path/to/schema.graphql User posts "component showing list of posts with title and author username"

This command will generate the component

Config params

prompt_info - you can add 'baked' additional prompt information into your axolotl config file

{
    "frontend_prompt_info": "use shadcn components: Input, Button, Select",
    "agent_model": "gpt-4.1-nano"
}

Generate code

You can also generate typescript code - universal one for webhooks utility functions etc

npx axolotl cai schema.graphql "stripe webhook to handle payments"