go-cerebras

Unofficial Go client for the Cerebras Inference API, generated with oapi-codegen. cerebras.NewClientWithResponses("https://api.cerebras.ai", cerebras.WithRequestEditorFn(...)) from github.com/psyb0t/go-cerebras gives typed methods for all 7 paths of the spec, chat completions (ChatV1ChatCompletionsPost), text completions (CompletionsV1CompletionsPost), model listing and lookup, and TCP warming. Union schemas come with As.../From... helpers. Use when Go code needs to call Cerebras inference, or already imports go-cerebras.

Typed Go client for the Cerebras Inference API. Every operation in the spec is a method on ClientWithResponses, named after its operation ID.

For install, the method list, unions and error handling in detail, see references/setup.md.

Security & safety

• Read the API key from an environment variable such as CEREBRAS_API_KEY. Never hardcode it, commit it or log it, and never print request headers.
• Requests go only to the base URL passed to NewClientWithResponses. Prompts sent through the client reach Cerebras.

When to use

• Go code calls Cerebras for chat or text completions, or lists the models it serves.
• The project already imports github.com/psyb0t/go-cerebras.

When NOT to use

• You need streaming. The generated client returns the whole body; for stream: true read HTTPResponse.Body from the plain Client methods yourself.
• The code is not Go. Use Cerebras's own SDKs.

Usage

go

import (
	"context"
	"fmt"
	"net/http"
	"os"

	"github.com/psyb0t/go-cerebras"
)

auth := cerebras.WithRequestEditorFn(func(_ context.Context, req *http.Request) error {
	req.Header.Set("Authorization", "Bearer "+os.Getenv("CEREBRAS_API_KEY"))

	return nil
})

client, err := cerebras.NewClientWithResponses("https://api.cerebras.ai", auth)
if err != nil {
	return err
}

var content cerebras.UserMessageRequest_Content
if err := content.FromUserMessageRequestContent0("Say hi"); err != nil {
	return err
}

role := cerebras.UserMessageRequestRoleUser

var msg cerebras.ChatCompletionRequest_Messages_Item
if err := msg.FromUserMessageRequest(cerebras.UserMessageRequest{Content: content, Role: &role}); err != nil {
	return err
}

res, err := client.ChatV1ChatCompletionsPostWithResponse(ctx,
	&cerebras.ChatV1ChatCompletionsPostParams{},
	cerebras.ChatCompletionRequest{Model: "<model-id>", Messages: []cerebras.ChatCompletionRequest_Messages_Item{msg}},
)
if err != nil {
	return err // transport error
}

if res.JSON200 == nil {
	return fmt.Errorf("chat: status %d: %s", res.StatusCode(), res.Body)
}

completion, err := res.JSON200.AsChatCompletionResponse()
if err != nil {
	return err
}

fmt.Println(*completion.Choices[0].Message.Content)

Listing models:

go

res, err := client.ListModelsV1ModelsGetWithResponse(ctx, &cerebras.ListModelsV1ModelsGetParams{})
if err != nil {
	return err
}

list, err := res.JSON200.AsModelMetadataList()

Check JSON200 for nil before reading it. A non-2xx status is not a Go error; the status is in StatusCode() and the raw body in Body.