> ## Documentation Index
> Fetch the complete documentation index at: https://docs.nekt.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Streamlit

> Build data apps in Streamlit with direct access to your data in Nekt.

Connect Streamlit to Nekt using the Nekt SDK and a scoped access token. The token grants read access to the specific tables you choose, and can be used across all cloud deployments.

|                  | Available |
| :--------------- | :-------- |
| **Nekt Express** | Yes       |
| **GCP**          | Yes       |
| **AWS**          | Yes       |

***

## Generate a token

<Steps>
  <Step title="Create new access token">
    In the Nekt platform, go to the [Access Tokens](https://app.nekt.ai/notebooks/access-tokens) page.

    Click **Create a token**.
  </Step>

  <Step title="Select tables">
    Select the tables you want to give access to. You can select tables from any layer of your Lakehouse.

    <Note>
      The generated token grants access only to the tables selected here. You can create multiple tokens to give access to different sets of tables — useful for scoping access per application or user.
    </Note>

    Click **Create token**.
  </Step>

  <Step title="Copy the token">
    A success message will confirm the token was created. Go to the [Access Tokens](https://app.nekt.ai/notebooks/access-tokens) page to view it and copy your token.
  </Step>
</Steps>

***

## Connect with the Nekt SDK

The Nekt SDK lets your Streamlit app load tables directly from your Lakehouse as Spark DataFrames.

### Project setup

Your Streamlit project needs three files alongside your app code:

**`requirements.txt`** — Python dependencies:

```txt theme={null}
streamlit
git+https://github.com/nektcom/nekt-sdk-py.git#egg=nekt-sdk
```

**`packages.txt`** — System packages required by the SDK's Spark runtime:

```txt theme={null}
openjdk-17-jdk
```

<Note>
  The `packages.txt` file is used by Streamlit Community Cloud to install system-level dependencies. If you deploy elsewhere, install Java 17 through your platform's package manager or Docker image.
</Note>

### Store the token as a secret

Never hardcode your token. Use [Streamlit secrets management](https://docs.streamlit.io/deploy/streamlit-community-cloud/deploy-your-app/secrets-management) to store it securely.

Create a `.streamlit/secrets.toml` file in your project root:

```toml theme={null}
DATA_ACCESS_TOKEN = "your-nekt-token-here"
```

<Warning>
  Add `.streamlit/secrets.toml` to your `.gitignore` to keep the token out of version control.
</Warning>

When deploying to Streamlit Community Cloud, add the same key-value pair in your app's **Secrets** settings.

### Load data in your app

Initialize the SDK with the token from secrets, then call `nekt.load_table()` to fetch tables from your Lakehouse:

```python theme={null}
import streamlit as st
import nekt

nekt.data_access_token = st.secrets["DATA_ACCESS_TOKEN"]

df = nekt.load_table(layer_name="Raw", table_name="orders")

st.dataframe(df.toPandas())
```

`load_table` returns a Spark DataFrame. Call `.toPandas()` to convert it for use with Streamlit's display components.

You can also select specific columns before converting:

```python theme={null}
df = nekt.load_table(
    layer_name="Raw",
    table_name="orders"
).select("id", "customer_name", "total")

st.dataframe(df.toPandas())
```

### Full example

```python theme={null}
import streamlit as st
import nekt

nekt.data_access_token = st.secrets["DATA_ACCESS_TOKEN"]

st.title("Sales Dashboard")

orders_df = nekt.load_table(
    layer_name="Raw",
    table_name="orders"
).select("id", "customer_name", "total")

st.metric("Total orders", orders_df.count())
st.dataframe(orders_df.toPandas())
```

<Tip>
  For a complete working project you can fork and deploy, see the [streamlit-demo repository](https://github.com/nektcom/streamlit-demo).
</Tip>

***

## Deploy to Streamlit Community Cloud

<Steps>
  <Step title="Push your code to GitHub">
    Your repository should contain at least `streamlit_app.py`, `requirements.txt`, and `packages.txt`. Do not commit `.streamlit/secrets.toml`.
  </Step>

  <Step title="Create a new app in Streamlit Community Cloud">
    Go to [share.streamlit.io](https://share.streamlit.io), click **New app**, and select your repository, branch, and main file path.
  </Step>

  <Step title="Add your Nekt token to Secrets">
    In the app's **Advanced settings**, paste your secret:

    ```toml theme={null}
    DATA_ACCESS_TOKEN = "your-nekt-token-here"
    ```
  </Step>

  <Step title="Deploy">
    Click **Deploy**. Streamlit will install the system packages from `packages.txt`, the Python dependencies from `requirements.txt`, and start your app.
  </Step>
</Steps>

Once deployed, you're ready to go!

***

## SDK reference

The following SDK methods are available for use in your Streamlit app. See the full [SDK documentation](/notebooks/sdk) for details on all methods.

| Method                                    | Description                                           |
| :---------------------------------------- | :---------------------------------------------------- |
| `nekt.load_table(layer_name, table_name)` | Load a table from your Lakehouse as a Spark DataFrame |
| `nekt.get_spark_session()`                | Access the Spark session for advanced operations      |

***

## Need help?

Contact our support team if you encounter issues during setup.