Generative AI

Emulates the managed generative-AI services — the model catalogs, deployments, guardrails, and agents of AWS Bedrock and Azure OpenAI (Microsoft.CognitiveServices) — plus a runtime that answers inference calls. Each ships a control plane you provision against and a deterministic runtime that echoes your prompt back inside the model-native response envelope, so there are no GPUs, accounts, or per-token costs involved.

Reach for it in tests when your code invokes a model, deploys a model version, or orchestrates a retrieval-augmented agent — so you can exercise wiring, retries, and error handling without a live model. Because responses are deterministic, assertions stay stable across runs. For the training/endpoint/job side of AI — SageMaker, Vertex AI, and Azure ML — see Machine Learning.

Drive it with the real SDK#

Drop the SDK-compat server in front of cloudemu and point the real runtime client at it — this exercises the actual invoke path, request signing, and response parsing your production code uses:

import (
    "github.com/aws/aws-sdk-go-v2/service/bedrockruntime"
    "github.com/stackshy/cloudemu/v2"
    awsserver "github.com/stackshy/cloudemu/v2/server/aws"
)

cloud := cloudemu.NewAWS()
ts := httptest.NewServer(awsserver.New(awsserver.Drivers{
    Bedrock:             cloud.Bedrock,
    BedrockAgent:        cloud.BedrockAgent,
    BedrockAgentRuntime: cloud.BedrockAgentRuntime,
}))
defer ts.Close()

rt := bedrockruntime.NewFromConfig(cfg, func(o *bedrockruntime.Options) {
    o.BaseEndpoint = aws.String(ts.URL)
})

rt.InvokeModel(ctx, &bedrockruntime.InvokeModelInput{
    ModelId: aws.String("anthropic.claude-3-sonnet-20240229-v1:0"),
    Body:    []byte(`{"messages":[{"role":"user","content":"hello"}]}`),
})

Azure OpenAI works the same way: register the mock as CognitiveServices (ARM) and AzureAIDataPlane (inference) on azureserver.Drivers, then point armcognitiveservices at the ARM endpoint and azopenai at the *.openai.azure.com data plane. See the SDK-Compat Server page for the Azure TLS setup.

Call the driver directly#

When you don't need the SDK round-trip, call the driver. ListFoundationModels reads the seeded catalog, InvokeModel runs the deterministic runtime over a model-native body, and the agent runtime answers RAG calls:

import bedrockdriver "github.com/stackshy/cloudemu/v2/services/bedrock/driver"

models, _ := aws.Bedrock.ListFoundationModels(ctx)

out, _ := aws.Bedrock.InvokeModel(ctx, bedrockdriver.InvokeModelInput{
    ModelID: "anthropic.claude-3-sonnet-20240229-v1:0",
    Body:    []byte(`{"prompt":"hello"}`),
})

// Bedrock Agents: build a knowledge base + agent, then answer a RAG query.
kb, _ := aws.BedrockAgent.CreateKnowledgeBase(ctx, /* KnowledgeBaseConfig */)
ans, _ := aws.BedrockAgentRuntime.RetrieveAndGenerate(ctx, /* RetrieveAndGenerateInput */)

On the Azure side, azure.AI provisions AI Services accounts and model deployments, then serves inference:

acct, _ := azure.AI.CreateAccount(ctx, /* AccountConfig */)         // Microsoft.CognitiveServices/accounts
dep, _  := azure.AI.CreateDeployment(ctx, /* DeploymentConfig */)   // gpt-4o, text-embedding-3-*, …

Behavior & fidelity#

SDK-compat — Live#

Real bedrock, bedrock-agent, and Azure OpenAI (armcognitiveservices + azopenai) clients drive it end-to-end:

See SDK-Compat for the full per-operation list.