You can use Opik through Comet's Managed Cloud offering or you can self-host Opik on your own infrastructure. When choosing to self-host Opik, you get access to all Opik features including tracing, evaluation, etc but without user management features.

If you choose to self-host Opik, you can choose between two deployment options:

Local installation: Perfect to get started but not production-ready.
Kubernetes installation: Production ready Opik platform that runs on a Kubernetes cluster.

Getting started

If you would like to try out Opik locally, we recommend using our Local installation script. Assuming you have git and docker installed, you can get started with just one line of code:

<Tabs> <Tab value="Linux / Mac" title="Linux / Mac"> ```bash # Clone the Opik repository git clone https://github.com/comet-ml/opik.git

Navigate to the opik folder

cd opik

Start the Opik platform

./opik.sh


</Tab>
<Tab value="Windows" title="Windows">

```powershell
# Clone the Opik repository
git clone https://github.com/comet-ml/opik.git

# Navigate to the opik folder
cd opik

# Start the Opik platform
powershell -ExecutionPolicy ByPass -c ".\opik.ps1"

</Tab> </Tabs>

Opik will now be available at

<a href="http://localhost:5173" target="_blank"> http://localhost:5173 </a> and all traces logged from your local machine will be logged to this local Opik instance. In order for traces and other data to be logged to your Opik instance, you need to make sure that the Opik Python SDK is configured to point to the Opik server you just started. You can do this by running the following command:

bash

# Configure the Python SDK to point to the local Opik platform
export OPIK_BASE_URL=http://localhost:5173/api

or in Python:

python

import os

os.environ["OPIK_BASE_URL"] = "http://localhost:5173/api"

To learn more about how to manage you local Opik deployment, you can refer to our local deployment guide.

SDK version compatibility

For self-hosted deployments, we recommend keeping the Opik server and SDKs on the latest versions for best compatibility. The Python SDK (opik), TypeScript SDK (opik), and Opik Agent Optimizer SDK (opik-optimizer) are designed to work with the latest Opik server release. Older SDK versions may not support newer server features, and newer SDK versions may require a minimum server version. Check the changelog for version-specific compatibility notes.

Advanced deployment options

If you would like to deploy Opik on a Kubernetes cluster, we recommend following our Kubernetes deployment guide here.

Comet managed deployments

The Opik platform is being developed and maintained by the Comet team. If you are looking for a managed deployment solution, feel free to reach out to the Comet team at [email protected] or visit the Comet website to learn more.