Quick Start Guide#

The fastest way to get productive on Apothic today is:

create or sign in to your Apothic account
create an API key
fund your account balance if you plan to run billable workloads
install apothic-client
deploy your first app

Choose your starting point#

There are three common entry points:

Apothic Client: Python SDK and CLI for apps, functions, services, stateful classes, sandboxes, volumes, queues, dictionaries, and operator workflows
Dashboard: website flows for account, billing, managed deployments, and admin settings
SMCP: dynamic MCP sessions for scriptable tool environments

If you want programmable runtime apps and jobs, start with apothic-client.

Prerequisites#

an Apothic account
Python 3.11 or newer
uv or pip

Step 1: create an API key#

Get an API key from your Apothic account and export it locally:

export APOTHIC_API_KEY=your_apothic_api_key

Optional override:

export APOTHIC_BASE_URL=https://run.apothic.ai

Step 2: make sure your balance is funded#

Runtime usage is tied to your Apothic account balance.

Before you run billable workloads, add funds from the website billing flow:

open /dashboard/billing
use Add Funds
complete the top-up checkout

Step 3: install `apothic-client`#

With uv:

uv add apothic

With pip:

pip install apothic

Step 4: define your first app#

Create hello_app.py:

from apothic import App, RemoteFunction

app = App("hello-world")


@app.function(cpu=1, memory_mb=512, timeout_s=60)
def greet(name: str) -> str:
    return f"hello, {name}"


@app.local_entrypoint()
def main() -> None:
    deployment_id = app.deploy()
    remote = RemoteFunction.from_name("hello-world", "greet")
    print("deployment:", deployment_id)
    print(remote.remote("world"))

Step 5: deploy and run it#

Deploy from the CLI:

apothic deploy hello_app.py

Or invoke through the local entrypoint:

python hello_app.py

Run the function from the CLI:

apothic run greet \
  --app-name hello-world \
  --payload '{"args":["world"],"kwargs":{}}' \
  --watch

Step 6: explore the next layer#

After the first function works, the next useful capabilities are:

background jobs and retries
service deployments with @app.endpoint(...) or @app.asgi(...)
stateful resources with @app.cls(...) and app.sandbox(...)
shared worker coordination with Queue and Dict
runtime-managed storage with portable named volumes, vast_local, and CloudBucketMount(...)
secrets and structured capacity filters through offer_filters
direct base-image execution, with managed-CPython fallback for Python-less Linux base images when needed
watch streams for jobs, apps, deployments, queues, and dictionaries
generated implementations with @apothic.liv(...)