Amber

Local-first Azure emulator — Build, test, and iterate on Azure services without an Azure account.

Amber emulates 86 Azure service-family surfaces across 17 categories, all running locally with zero core runtime dependencies. Every operation records billing usage so you can estimate costs before deploying to real Azure.

Amber is a developer test double, not Microsoft Azure. It provides useful local behavior, ARM-shaped resource IDs, region and zone metadata, billing ledgers, and adapter hooks while documenting where cloud validation is still required.

Amber is inspired by the local-cloud developer experience pioneered by Floci for AWS, and brings that same kind of fast local workflow to Azure-shaped Python, CLI, and REST development.

Quick Start

from amber import AmberWorkspace

ws = AmberWorkspace.create(".amber")

# Blob Storage
ws.storage.blob.create_container("data")
ws.storage.blob.put_blob("data", "hello.txt", b"Hello, Azure!")

# Redis Cache
ws.redis.set(0, "session:1", "active")
assert ws.redis.get(0, "session:1") == "active"

# Cosmos DB
ws.cosmosdb.create_account("myaccount")
ws.cosmosdb.create_database("myaccount", "mydb")
ws.cosmosdb.create_container("myaccount", "mydb", "items", "/category")
ws.cosmosdb.create_document("myaccount", "mydb", "items", {
    "id": "doc1", "category": "fruit", "name": "apple"
})

# Service Bus messaging
ws.servicebus.create_queue("orders")
ws.servicebus.send_to_queue("orders", "new-order-123")

# Check billing
for key, usage in sorted(ws.billing.summary().items()):
    print(f"{key}: {usage.quantity} {usage.unit}")

ws.close()

Installation

pip install amber-cloud

Amber is published on PyPI as amber-cloud, while the import package and CLI are both amber.

Documentation

• Documentation: https://vishwa-docs.github.io/amber/
• PyPI: https://pypi.org/project/amber-cloud/
• Issues and roadmap: https://github.com/Vishwa-docs/amber/issues

Local Azure Workflows

amber doctor
amber az group create --name apps --location eastus --tags env=dev
amber az group list
eval "$(amber env --endpoint http://127.0.0.1:8000)"

For an optional ARM-shaped local endpoint:

pip install "amber-cloud[server]"
python -m amber.server --workspace .amber --host 127.0.0.1 --port 8000

Why Amber

• Run Azure-shaped integration tests locally without cloud accounts or shared test subscriptions.
• Keep application code close to ARM-shaped IDs, locations, tags, resource groups, and service-family names.
• Prototype lift-and-shift flows before replacing local adapters with real Azure SDK clients or deployment pipelines.
• Give students, teams, and open-source projects a fast local playground for Azure concepts.

Service Coverage (86 services)

Public Python API

Amber is packaged as amber-cloud and imported as amber:

from amber import AmberWorkspace
from amber.core import ServiceRegistry
from amber.cosmosdb import CosmosDBService
from amber.storage import BlobStorageService, QueueStorageService
from amber.networking import VNetService

The subpackages intentionally mirror Azure service families so tests and prototypes can move from local Amber code toward Azure SDK or deployment workflows with minimal conceptual translation. See the Sphinx documentation's Microsoft References page for the official Microsoft docs that each local service surface maps to.

Architecture

• SQLite persistence with WAL mode for concurrent access
• Content-addressed blob storage with SHA-256 deduplication
• Lazy service initialization — only loaded when accessed
• Thread-safe workspace with double-checked locking
• Zero external dependencies — pure Python stdlib

Commands

Go Core

The Go runtime and CLI remain alongside the Python package:

go test ./...
go run ./cmd/amber --workspace .amber doctor
go run ./cmd/amber --workspace .amber status

License

Apache 2.0. Azure, Microsoft Azure, and related service names are trademarks of Microsoft Corporation. Amber is independent and is not affiliated with, endorsed by, or sponsored by Microsoft Corporation.