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

# SDK Reference

> All functions and classes exposed in the top layer of the SDK

These are the functions and classes available with simply `import agentops`

## Functions

***

### `init()`

The first element of AgentOps is always calling .init()

**Parameters**:

* `api_key` (str, optional): API Key for AgentOps services. If not provided, the key will be read from the `AGENTOPS_API_KEY` environment variable.
* `endpoint` (str, optional): The endpoint for the AgentOps service. Defaults to '[https://api.agentops.ai](https://api.agentops.ai)'.
* `max_wait_time` (int, optional): The maximum time to wait in milliseconds before flushing the queue. Defaults to 30,000 (30 seconds).
* `max_queue_size` (int, optional): The maximum size of the event queue. Defaults to 100.
* `tags` (List\[str], optional): Tags for the sessions for grouping or sorting (e.g., \["GPT-4"]).
* `override` (bool, optional): \[Deprecated] Use `instrument_llm_calls` instead to decide whether to instrument LLM calls and emit LLMEvents.
* `instrument_llm_calls` (bool): Whether to instrument LLM calls and emit LLMEvents.
* `auto_start_session` (bool): Whether to start a session automatically when the client is created. You may wish to delay starting a session in order to do additional setup or starting a session on a child process.
* `inherited_session_id` (str, optional): When creating the client, passing in this value will connect the client to an existing session. This is useful when having separate processes contribute to the same session.
* `skip_auto_end_session` (bool, optional): If you are using a framework such as Crew, the framework can decide when to halt execution. Setting this parameter to true will not end your agentops session when this happens.

**Returns**:

* Session ID of the current or new session.

***

### `start_session()`

Start a new [Session](/v1/concepts/sessions) for recording events.

**Parameters**:

* `tags` (List\[str], optional): Tags for the session.
* `config` ([Config](#config), optional): Client configuration object.
* `inherited_session_id` (str, optional): Session ID to continue from an existing session.

**Returns**:

* Session ID of the started session.

***

### `end_session()`

Ends the current session with the AgentOps service.

**Parameters**:

* `end_state` (str): The final state of the session. Options: Success, Fail, or Indeterminate.
* `end_state_reason` (str, optional): The reason for ending the session.
* `video` (str, optional): URL to a video recording of the session.

***

### `record()`

Record an event with the AgentOps service.

**Parameters**:

* `event` (Union\[Event, ErrorEvent]): The event to record.

***

### `get_api_key()`

Retrieve the API key used by the client.

**Returns**:

* API key as a string.

***

## Types

***

### `Config`

Stores the configuration settings for AgentOps clients.

**Parameters**:

* `api_key` (str, optional): API Key for AgentOps services. If not provided, the key will be read from the `AGENTOPS_API_KEY` environment variable. If no key is found, a `ConfigurationError` is raised.
* `endpoint` (str, optional): The endpoint for the AgentOps service. If not provided, the endpoint will be read from the `AGENTOPS_API_ENDPOINT` environment variable. Defaults to '[https://api.agentops.ai](https://api.agentops.ai)'.
* `max_wait_time` (int, optional): The maximum time to wait in milliseconds before flushing the queue. Defaults to 30000.
* `max_queue_size` (int, optional): The maximum size of the event queue. Defaults to 100.

**Properties**

* **api\_key** (str): Get or set the API Key for AgentOps services.
* **endpoint** (str): Get or set the endpoint for the AgentOps service.
* **max\_wait\_time** (int): Get or set the maximum wait time in milliseconds before flushing the queue.
* **max\_queue\_size** (int): Get or set the maximum size of the event queue.

***

### `LangchainCallbackHandler`

[Reference](/v1/usage/langchain-callback-handler)

This callback handler is intended to be used as an option in place of AgentOps auto-instrumenting. This is only useful
when using LangChain as your LLM calling library.

<script type="module" src="/scripts/github_stars.js" />

<script type="module" src="/scripts/scroll-img-fadein-animation.js" />

<script type="module" src="/scripts/button_heartbeat_animation.js" />

<script type="module" src="/scripts/adjust_api_dynamically.js" />