@rustic-ai/api-client

TypeScript/axios client for the Rustic AI HTTP API.

The client is generated from the OpenAPI specification (forge-go/api/openapi/) with OpenAPI Generator (typescript-axios). Do not edit src/ by hand — it is overwritten on every regeneration.

Install

npm install @rustic-ai/api-client axios

axios is a runtime dependency and is installed automatically.

Usage

import { Configuration, GuildsApi, CatalogApi } from '@rustic-ai/api-client'

const config = new Configuration({ basePath: 'http://localhost:8880' })

const guilds = new GuildsApi(config)
const { data } = await guilds.getGuildDetailsById({ guildId: 'my-guild' })

Model interfaces (BlueprintCreate, AgentSpecInput, GuildSpec, …) are exported from the package root:

import type { BlueprintDetailsResponse, GuildStatus } from '@rustic-ai/api-client'

Regenerating

The client is generated directly from the OpenAPI 3.1 document (forge-go/api/openapi/openapi.json).

npm install        # once
npm run generate   # regenerate src/ from the spec
npm run build      # type-check and emit dist/

Generator settings live in openapitools.json (pinned generator version, input spec, options). Files that must not be produced inside src/ are listed in src/.openapi-generator-ignore.

The generate script also applies a checked compatibility annotation and return cast to createRequestFunction. Axios 1.19 and newer expose a private unique-symbol type through inference, so declaration builds require this generated function to retain its explicit public Promise<R> contract. The patch script fails if the upstream generator output changes, requiring the override to be reviewed.