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

# Agentic Functions

### `@agentic`

{/* START */}

Decorator for creating LLM-implemented functions.
This decorator is used on functions that will be implemented by an LLM.
The decorated function should have a descriptive docstring but an empty
body (containing only `...`).

{/* @signature agentic customer_sdk/python/src/agentica/decorator/decorator.py */}

```python theme={null}
def agentic[**I, O](
    *scope_defined: Any,
    scope: dict[str, Any] | None = None,
    system: str | None = None,
    premise: str | None = None,
    mcp: str | None = None,
    persist: bool = False,
    model: ModelId = DEFAULT_AGENT_MODEL,
    listener: Callable[[], AgentListener] | DefaultAgentListener | None = DEFAULT_AGENT_LISTENER,
    max_tokens: int | MaxTokens = MaxTokens.default(),
    reasoning_effort: ReasoningEffort | None = None,
    cache_ttl: CacheTTL | None = None,
) -> Callable[[Callable[I, O]], Callable[I, O]]:
    ...
```

**Parameters**

{/* START */}

<ParamField body="*scope_defined" type="Any">
  A list of runtime resources as in `scope`. The names of the resources are not specified explicitly and are instead derived automatically from the resources themselves.
  `scope` and `scope_defined` can be used together to specify resources with both explicit and implicit names. The names can't be repeated between the two.
  Example:

  ```
  @agentic(my_func, db_connection, cache) # the same as @agentic(scope={"my_func": my_func, "db_connection": db_connection, "cache": cache})
  async def my_function():
      ...
  ```
</ParamField>

<ParamField body="scope" type="dict[str, Any]">
  A dictionary of names mapped to runtime resources that are in scope and which may be used during the execution of the agentic function.
  Resources in scope may be arbitrary Python functions, methods, objects, iterators, types or any other Python value.
</ParamField>

<ParamField body="system" type="str | None">
  An optional system prompt for the agentic function.
  This will be the system prompt of all invocations of this agentic function.
  This argument cannot be provided along with the `premise` argument.
</ParamField>

<ParamField body="premise" type="str | None">
  An optional premise for the function.
  This will be attached to the system prompt of all invocations of this agentic function.
  This argument cannot be provided along with the `system` argument.
</ParamField>

<ParamField body="mcp" type="str | None">
  The string of a path to a .json file representing an MCP configuration.
  Any servers and/or tools of servers outlined in the config can be used during the execution of the agentic function.
</ParamField>

<ParamField body="persist" type="bool">
  Whether to persist the function state/history between calls.
</ParamField>

<ParamField body="model" type="str">
  The model used to execute the agentic function.
  Any OpenRouter model slug is supported.
</ParamField>

<ParamField body="listener" type="Callable[[], AgentListener] | None">
  Optional listener constructor for logging the agentic function's activity and chat history.
  If None, no listener will be used.
</ParamField>

<ParamField body="max_tokens" type="int | MaxTokens">
  When an integer is supplied, this is the maximum number of tokens for an invocation.
  For more fine-grained control, a `MaxTokens` object may be passed.
</ParamField>

<ParamField body="reasoning_effort" type="'minimal' | 'low' | 'medium' | 'high' | 'xhigh' | None">
  Constrains thinking budget on reasoning models which support it (gpt 5.2, sonnet 4.5, gemini 3, etc...)
  Higher values use more reasoning tokens but may produce better results.
  If None, uses the model's default reasoning effort.
</ParamField>

<ParamField body="cache_ttl" type="'5m' | '1h' | None">
  Controls how long Anthropic prompt caching entries persist.
  Only used for Anthropic models; ignored for other providers.
</ParamField>

<Warning>
  The `system` argument cannot be provided along with the `premise` argument, and vice versa.
</Warning>

<Accordion title="MCP Configuration Fields">
  <ParamField path="command" type="string" required>
    The executable command to run the MCP server. This should be an absolute path or a command available in the system `PATH`.

    Example:

    ```
    "python"
    ```
  </ParamField>

  <ParamField path="args" type="array of string">
    An array of command-line arguments passed to the server executable. Arguments are passed in order.

    Example:

    ```json theme={null}
    ["server.py", "--verbose", "--port", "8080"]
    ```
  </ParamField>

  <ParamField path="env" type="object">
    An object containing environment variables to set when launching the server. All values must be strings.

    Example:

    ```json theme={null}
    {
        "API_KEY": "secret-key",
        "PORT": "8080"
    }
    ```
  </ParamField>
</Accordion>

<Note>
  The default model is `openai/gpt-4.1`.
</Note>

<Note>
  The default agent listener is the `StandardListener`, but can be changed for all agents and agentic functions in the current scope with `set_default_agent_listener`.
  If a context-specific logger is used in the current scope, the logger will be added to the listener: if the listener is `None`, then the listener will be set to:

  * the default agent listener, if it is not `None`, or
  * the `StandardListener`, if the default agent listener is `None`

  For more information on the `StandardListener` and the listener hierarchy, see [here](/concepts/logging_and_streaming/python).
</Note>

{/* START */}

<ResponseField name="Returns" type="Callable[[Callable[I, O]], Callable[I, O]]">
  The decorated function that will be implemented by the LLM.
</ResponseField>