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

# Langflow

> Use TokenLab models from Langflow's OpenAI component

## Overview

Langflow's **OpenAI** component includes an **OpenAI API Base** field. Point that field at TokenLab's OpenAI-compatible `/v1` endpoint and use TokenLab model IDs.

<Note>
  **Type**: Low-code AI Workflow Builder

  **Primary Path**: OpenAI-compatible chat completions

  **Support Confidence**: Supported with scope limits
</Note>

## OpenAI Component

1. Add an **OpenAI** language model component.
2. Set **OpenAI API Key** to your TokenLab key.
3. Expand advanced settings.
4. Set **OpenAI API Base** to `https://api.tokenlab.sh/v1`.
5. Set **Model Name** to a TokenLab model ID.
6. Connect the model output to your agent, chain, or workflow.

## Values

| Field           | Value                        |
| --------------- | ---------------------------- |
| OpenAI API Base | `https://api.tokenlab.sh/v1` |
| OpenAI API Key  | `sk-your-tokenlab-key`       |
| Model Name      | `gpt-5.4-mini`               |

## Recommended Models

* `gpt-5.4-mini` for most workflow tests
* `claude-sonnet-5` for stronger writing and reasoning
* `gemini-3.5-flash` for fast multimodal-friendly workflows
* `deepseek-v4-flash` for fast coding and reasoning iterations

Use Langflow's LiteLLM Proxy component if your deployment already centralizes provider routing through a LiteLLM proxy.