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

# Link a Weave prompt to a collection

Track, version, and manage your AI prompts in the W\&B Registry by linking Weave prompt objects to a collection. *Linking* a prompt object to a [collection](/models/registry/create_collection) moves the prompt version from a [private, project-level scope to a shared, organization-level scope](/models/registry/create_registry#visibility-types).

<Note>
  Before you start, check the following:

  * The object types that collection permits. For more information about collection types, see "Collection types" within [Create a collection](./create_collection).
  * The registry that the collection belongs to already exists. To check that the registry exists, navigate to the [Registry App and search for](/models/registry/search_registry) the name of the registry.
</Note>

To track and manage your AI prompts, use the Weave Python SDK. The Weave Python SDK enables you to create, publish, and link prompt objects to a collection in a registry.

The following steps outline how to link a prompt to a collection in a registry:

1. Initialize a Weave client with `weave.init()`.
2. Create a Weave prompt object with a built-in prompt class such as [`weave.StringPrompt()`](/weave/reference/python-sdk#class-stringprompt).
   <Tip>
     `weave.StringPrompt` is a built-in prompt class for single-string prompts, such as a system prompt, user prompt, or reusable text template. For more information about creating prompt objects in Weave, and supported prompt types, see [Create prompt objects](/weave/guides/core-types/prompts).
   </Tip>
3. Publish the prompt to Weave with [`weave.publish()`](/weave/reference/python-sdk#method-publish).
4. Link the published prompt version to a collection in a registry with [`weave_client.link_prompt_to_registry()`](/weave/reference/python-sdk/trace/weave_client#method-link_prompt_to_registry). Pass the destination collection path to the `target_path` parameter. The path consists of the `wandb-registry-` prefix, the registry name, and the collection name:
   ```text theme={null}
   wandb-registry-{REGISTRY_NAME}/{COLLECTION_NAME}
   ```

Copy and paste the following code snippet into your Python script, replacing the placeholder values enclosed in `<>` with your own:

```python theme={null}
import weave

# Initialize Weave client
weave_client = weave.init(project="<project>")

prompt = "<You are a helpful assistant...>"

# Publish the prompt object to Weave
weave_prompt_ref = weave.publish(
    obj=weave.StringPrompt(prompt),
    name="<prompt_name>",
)

# Specify the collection and registry to link the prompt version to
REGISTRY_NAME = "<registry_name>"
COLLECTION_NAME = "<collection_name>"
registry_target = f"wandb-registry-{REGISTRY_NAME}/{COLLECTION_NAME}"

# Link the published prompt version to a collection in the specified registry
weave_client.link_prompt_to_registry(
    prompt=weave_prompt_ref,
    target_path=registry_target,
)
```

## View linked prompts in a registry

1. Navigate to the W\&B Registry.
2. Select the name of the registry that you linked the prompt to.
3. Select the name of the collection.
4. From the list of prompt versions, select the version you want to access. Version numbers are incrementally assigned to each linked prompt version starting with `v0`.
5. To view details about a prompt version, click the version. From the tabs in this page, you can view that version's overview, prompt, and use.

## Use a prompt linked to a registry

Use the `weave.get()` method to retrieve a prompt object linked to a registry. You can fetch a prompt by its alias or version number. The following code snippet shows how to fetch a prompt by its alias and version number.

To fetch a prompt object by its alias, use the following URI path:

```text theme={null}
registry:///{ORG}/{REGISTRY_NAME}/{COLLECTION_NAME}:{VERSION-OR-ALIAS}
```

Replace the placeholder values enclosed in `<>` with your own:

<Tabs>
  <Tab title="Fetch by alias">
    ```python theme={null}
    # Fetch by alias
    URI_path = f"registry:///{ORG}/{REGISTRY_NAME}/{COLLECTION_NAME}:{ALIAS}"

    # Fetch by alias
    prompt = weave.get(URI_path)

    # Use the prompt
    print(prompt.content)
    ```
  </Tab>

  <Tab title="Fetch by version">
    ```python theme={null}
    # Fetch by version
    URI_path = f"registry:///{ORG}/{REGISTRY_NAME}/{COLLECTION_NAME}:{VERSION}"

    # Fetch by version
    prompt = weave.get(URI_path)

    # Use the prompt
    print(prompt.content)
    ```
  </Tab>
</Tabs>

W\&B autogenerates the URI path and sample code to fetch your prompt object. To view the sample code:

1. Navigate to the W\&B Registry.
2. Select the name of the registry that you linked the prompt to.
3. Select the name of the collection.
4. From the list of prompt versions, select the version you want to access.
5. Select the **Use** tab to view the sample code to fetch your prompt by alias or version number.