Function App Emulator

Preview

During Public Preview you can use the Team plan for free, and get an early-adopter discount at GA. Sign in

Function App Emulator

Available on Starter Standard Team Compare plans →

Azure Functions is a serverless compute service - running your code in response to events, such as HTTP requests, timers or messages, without you managing the servers it runs on.

Locally supports both provisioning Function Apps and deploying and running your code - so your triggers fire from Locally's other emulators, and your bindings read and write real data.

When you provision a Function App in Locally, it's available at https://<app>.furnace.locally:5663 - and you can deploy to it using the Azure CLI, func azure functionapp publish, or anything else which deploys through Kudu.

For a step-by-step walkthrough, see Provisioning a Function App.

Plugins required

Locally v2026.10+ automatically installs these Resource Provider plugins when their Resource Providers are registered, but you can also install them via:

$ locally plugin install --name Microsoft.Storage
$ locally plugin install --name Microsoft.Web

Storage is needed because every Function App keeps its state in a Storage Account. The language workers come from Azure Functions Core Tools, so func needs to be on your PATH.

Finding the Emulator

Once a Function App has been created, its resource page within the Locally Dashboard shows its host name, runtime, app settings, and the functions loaded from the last deployment - and Open in emulator takes you to the emulator:

Screenshot of a Function App in the Locally Dashboard, showing its configuration, app settings and the functions loaded from the last deployment with their triggers and bindings

You can also open the emulator from the command line using locally function dashboard --name <app>.

The Function App Emulator

Within the Function App Emulator you can see every invocation with its trigger payload and bindings, along with the app's logs, deployments and keys:

Screenshot of the Function App Emulator built into Locally, listing invocations of an HTTP-triggered and a queue-triggered function with their status, duration and payload

Opening an invocation shows its response and the payloads of its output bindings - and HTTP invocations can be replayed:

Screenshot of a single invocation in the Function App Emulator, showing the HTTP response body, the queue output binding's payload and the Replay button

A Function App with Durable Functions also has a Durable tab, showing its task hub wherever it lives - in the app's storage account, in its SQL database, or on a Durable Task Scheduler. It lists the hub's orchestrations, which you can search and filter by status or name, and its entities with their state. Opening an orchestration shows its history, and lets you terminate, suspend, resume, raise an event, rewind, restart or purge it. You can also start an orchestration, or purge every finished one.

What's Supported

The Function App emulator supports:

HTTP Triggers Timer Triggers Queue & Blob Triggers Service Bus & Event Hubs Triggers Cosmos DB Change Feed Event Grid Triggers Durable Functions Durable Task Scheduler Managed Identity App Service Authentication

Functions run in the real Azure Functions language workers - Node, Python, .NET isolated, Java and PowerShell, plus custom handlers - using the language runtimes installed on your machine.

You can deploy using az functionapp deployment source config-zip, func azure functionapp publish or locally function deploy - and with SCM_DO_BUILD_DURING_DEPLOYMENT set, Node and Python dependencies are installed when you deploy.

HTTP triggers check authLevel using the function and host keys (?code= or x-functions-key), and App Service Authentication signs users in through the Directory Emulator.

Chaos Engineering (on the Team plan) can make a Function App throttle, fail, slow down or drop requests, so you can check how its callers cope - and hold back, skip or repeat a function's trigger. It can also freeze, crash or restart the app on a schedule. Function Apps have their own functions surface, separate from Web Apps.

Function Apps also work with Locally's other emulators, as you'd expect - for example:

  • Triggers and bindings talk to the Storage, Service Bus, Event Hubs, Cosmos DB and Event Grid Emulators - so a message sent to a queue really triggers your function.
  • Key Vault references in app settings are resolved from the Key Vault Emulator, using the app's managed identity.
  • Durable Functions' task hub is stored in the app's Storage Account - unless host.json selects the azureManaged storage provider, in which case orchestrations run on the Durable Task Scheduler its connection string names, or the mssql provider, in which case the task hub is kept in the app's SQL database on a SQL server within Locally, signed in to with a SQL login.
  • An app linked to an Application Insights component gets its connection string in APPLICATIONINSIGHTS_CONNECTION_STRING.
  • With a diagnostic setting, the host's logs and App Service Authentication decisions land in the Log Analytics Emulator's FunctionAppLogs and AppServiceAuthenticationLogs tables.

Examples

These end-to-end examples use the Function App emulator, and have everything you need to run them against Locally:

The locallybuild/examples repository also has Function Apps for each runtime, with queue, timer and HTTP triggers.

Differences from Azure

As of Locally v2026.11.03, the Function App emulator has the following differences from Azure:

  • Your functions run as processes on your machine rather than in a sandbox, using the language runtimes you have installed. Remote builds use them too, rather than Azure's build image.
  • In-process .NET apps (FUNCTIONS_WORKER_RUNTIME=dotnet) aren't supported - use the isolated worker model.
  • SignalR, SQL, Redis, Kafka and RabbitMQ bindings aren't available, so functions that use them are marked as in error.
  • The host doesn't record invocations in Application Insights itself, so only telemetry your code sends arrives.
  • Each app runs as a single instance, with no scale-out.
  • Deployment slots can be swapped, but traffic can't be routed between them.

Should you encounter any issues, please take a look at the troubleshooting section, reach out in our Discord or open an issue and we'll take a look.

Preview

During Public Preview you can use the Team plan for free, and get an early-adopter discount at GA. Sign in

A local cloud for you and your AI agents.

Your Azure infrastructure, running on your machine. Deploy in seconds, break things freely, and ship to Azure when you're ready.