|
| 1 | +--- |
| 2 | +title: BurnCloud |
| 3 | +slug: /bundles-burncloud |
| 4 | +description: Use BurnCloud's OpenAI-compatible models inside Langflow. |
| 5 | +--- |
| 6 | + |
| 7 | +import Icon from "@site/src/components/icon"; |
| 8 | + |
| 9 | +<Icon name="Blocks" aria-hidden="true" /> [**Bundles**](/components-bundle-components) contain custom components that support specific third-party integrations with Langflow. |
| 10 | + |
| 11 | +This page describes the components that are available in the **BurnCloud** bundle. |
| 12 | + |
| 13 | +For more information about BurnCloud features and API limits, see the [BurnCloud documentation](https://burncloud.com/). |
| 14 | + |
| 15 | +## BurnCloud text generation |
| 16 | + |
| 17 | +The **BurnCloud** component generates text through BurnCloud's OpenAI-compatible API gateway. It works with the same chat-completions schema as OpenAI, while letting you point to BurnCloud-hosted models or a private BurnCloud deployment. |
| 18 | + |
| 19 | +It can output either a **Model Response** ([`Message`](/data-types#message)) or a **Language Model** ([`LanguageModel`](/data-types#languagemodel)). The **Language Model** output is an instance of [`ChatOpenAI`](https://python.langchain.com/docs/integrations/chat/openai) configured to target BurnCloud's `/v1` endpoints. |
| 20 | + |
| 21 | +Use the **Language Model** output when you want to pass a BurnCloud model into another LLM-driven component, such as **Agent**, **Smart Function**, or **Prompt Template** components. |
| 22 | + |
| 23 | +### BurnCloud parameters |
| 24 | + |
| 25 | +import PartialParams from '@site/docs/_partial-hidden-params.mdx'; |
| 26 | + |
| 27 | +<PartialParams /> |
| 28 | + |
| 29 | +| Name | Type | Description | |
| 30 | +|------|------|-------------| |
| 31 | +| api_key | SecretString | Input parameter. Your BurnCloud API key. Required for authentication and for fetching the latest model list. | |
| 32 | +| base_url | String | Input parameter. Override the default `https://ai.burncloud.com` base URL if you host BurnCloud privately. The component appends `/v1` automatically when needed. (Advanced) | |
| 33 | +| model_name | String | Input parameter. BurnCloud model to use. Options update dynamically after you provide a valid API key and click <Icon name="RefreshCw" aria-hidden="true" /> **Refresh**. Defaults to `gpt-4o`. | |
| 34 | +| temperature | Float | Input parameter. Controls randomness. Range: `[0, 2]`. Defaults to `0.7`. (Advanced) | |
| 35 | +| top_p | Float | Input parameter. Alternative sampling control that limits the cumulative probability mass of candidate tokens. Range: `[0, 1]`. Defaults to `1.0`. (Advanced) | |
| 36 | +| max_tokens | Integer | Input parameter. Maximum number of tokens to generate. Leave empty to let BurnCloud decide. (Advanced) | |
| 37 | +| input_value | String | Input parameter. The prompt or chat content you want to send to the model. | |
| 38 | +| system_message | String | Input parameter. Sets the assistant's persona or high-level instructions. | |
| 39 | +| stream | Boolean | Input parameter. Streams partial results when enabled. | |
| 40 | +| output_parser | OutputParser | Input parameter. (Advanced) Parse the model response before passing it downstream. | |
| 41 | +| model_output | LanguageModel | Output parameter. A `ChatOpenAI` instance configured for BurnCloud. | |
| 42 | +| text_output | Message | Output parameter. The generated response from the selected BurnCloud model. | |
| 43 | + |
| 44 | +### Use BurnCloud in a flow |
| 45 | + |
| 46 | +1. Sign up for a [BurnCloud account](https://burncloud.com/) and generate an API key in the BurnCloud dashboard. |
| 47 | +2. In Langflow, open <Icon name="Blocks" aria-hidden="true" /> **Bundles** and drag the **BurnCloud** component into your flow. |
| 48 | +3. Paste your API key into **BurnCloud API Key**. Optionally set **Base URL** if your organization hosts BurnCloud privately. |
| 49 | +4. Click <Icon name="RefreshCw" aria-hidden="true" /> **Refresh** next to **Model** to load the latest BurnCloud-hosted model list, then pick the model you need. |
| 50 | +5. Configure sampling parameters such as **Temperature**, **Top P**, and **Max Output Tokens** (if required) along with your **System Message** and **Prompt**. |
| 51 | +6. Connect **Chat Input** → **BurnCloud** → **Chat Output** (or feed the **Language Model** output into downstream components like **Agent** or **Smart Function**). |
| 52 | +7. Click **Playground** to test requests and validate the connection before deploying the flow. |
0 commit comments