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

# Update Version

> Update an existing version.

## Returns

Returns the updated [Version](/concepts/product/version) object.

## Example

```python theme={"system"}
updated_version = galtea.versions.update(
    version_id=version_id,
    description="Iteration with a larger context window",
    facts={"model": "gpt-4o", "temperature": "0.7"},
)
```

## Parameters

<ResponseField name="version_id" type="string" required>
  ID of the version to update.
</ResponseField>

<ResponseField name="name" type="string" optional>
  New name for the version. A name of `v` followed by digits is reserved for a version number, so
  the platform accepts it only when it is this version's own label.
</ResponseField>

<ResponseField name="description" type="string" optional>
  New version description.
</ResponseField>

<ResponseField name="system_prompt" type="string" optional>
  New system prompt used for the version.
</ResponseField>

<ResponseField name="model_id" type="string" optional>
  New reference to a [model](/concepts/model) configured in the Galtea Platform.
</ResponseField>

<ResponseField name="dataset_description" type="string" optional>
  New description of the dataset used for training.
</ResponseField>

<ResponseField name="dataset_uri" type="string" optional>
  New URI to the dataset.
</ResponseField>

<ResponseField name="guardrails" type="string" optional>
  New configuration for safety guardrails.
</ResponseField>

<ResponseField name="parent_version_id" type="string" optional>
  New ID of the version this one was revised from. The parent must belong to the same product.
</ResponseField>

<ResponseField name="facts" type="dict[str, str]" optional>
  New [facts](/concepts/product/version#version-facts): the agent state this version represents, as
  a flat map of text to text. Two live versions of one product cannot hold the same facts.
</ResponseField>

<ResponseField name="otel_output_source_rule" type="dict" optional>
  New rule that recovers the user-visible reply from a non-LLM span.
</ResponseField>

<ResponseField name="otel_input_source_rule" type="dict" optional>
  New rule that recovers the user message from a non-LLM span.
</ResponseField>

<ResponseField name="conversation_endpoint_connection_id" type="string" optional>
  New [endpoint connection](/concepts/product/endpoint-connection) used for conversational
  interactions with your AI product.
</ResponseField>

<ResponseField name="initialization_endpoint_connection_id" type="string" optional>
  New endpoint connection executed before the conversation begins.
</ResponseField>

<ResponseField name="finalization_endpoint_connection_id" type="string" optional>
  New endpoint connection executed after the conversation ends.
</ResponseField>

<ResponseField name="phone_connection_id" type="string" optional>
  New phone connection used for voice interactions with your AI product.
</ResponseField>

<ResponseField name="web_rtc_connection_id" type="string" optional>
  New WebRTC connection used for voice interactions with your AI product.
</ResponseField>

<Note>
  All fields except `version_id` default to `PydanticUndefined` (from `pydantic_core`).
  Omit a field (or pass `PydanticUndefined`) to leave it unchanged.
  Pass `None` to explicitly clear an optional field.
  Pass a value to update it.
</Note>