DiscordJoin Community

An easy-to-use open-source developer framework for end-to-end AI observability.

Easily create full observability and automation layer for your AI Systems - from Data preprocessing to LLM monitoring.

Try Demo NowQuick Start
banner

Trusted by ML teams from

meta
amazon
microsoft
ibm
airbus
amd
mila
mit
integrations dots

Aim connects and integrates with your favorite tools

The Aim standard package comes with all integrations. If you'd like to modify the integration and make it custom, create a new integration package and share with others.

Learn more
integrations image
features dots

Core Aim capabilities

Aim comes with a set of fundamental features that allows to create complete logging apps for any kind of software - specifically AI Systems: LLM apps, AI pipelines etc.

Log anything from everywhere

AimStack

Declare any python object as a loggable type

  • Metrics
  • Images
  • Text
  • Custom Objects
  • Remote logging
Get Started
AimStack

Pythonic low-code UI

AimStack

Create custom UI for your logs using Aim’s low-code pythonic UI-builder

  • Full UI Kit
  • Full Data type compatibility
  • Extensible and customizable
Get Started
AimStack

Modular ecosystem of packages

AimStack

Create an end-to-end custom logging app and run by Aim.

  • Distributed as python packages
  • Composable
  • UI + Automations + Logging
Get Started
AimStack

Log Automationscoming soon

Automate over your logs and connect observability to the rest of your toolstack.

  • Actions
  • Callbacks
  • Beats

Connect Artifactscoming soon

Connect Artifacts to your logs.

  • Versioning
  • Retrieval
top-bg

Use Aim in any environment. Get started with Aim in a few steps

  • It’s very easy to get started with Aim.

  • This example logs a basic ML training run.

  • Aim logs record sequences as part of containers. In this case the Standard Package Run container.

  • Start Aim Server and Aim UI.

  • You are ready to go!! 🚀

Quick Start

1. Install Aim in your environment

pip install aim

2. Integrate Aim with your code


from aimstack.base import Run, Metric

# Create a run
run = Run()

run['hparams'] = {
    'lr': 0.001,
    'batch_size': 32
}

# Create a metric
metric = Metric(run, name='loss', context={'epoch': 1})

for i in range(1000):
      metric.track(i, epoch=1)

3. Start Aim server and Aim UI

aim server
aim ui
Quick Start
bottom-bg
demos dots

Try Aim in action with live demos

Check out live Aim demos NOW to see it in action.

Subscribe to Our Newsletter

Subscribe to our newsletter to receive regular updates about our latest releases, tutorials and blog posts!

Subscribe