Skip to main content
Langflow is a low-code visual builder for multi-agent and RAG applications. This guide shows how to connect SambaNova’s fast inference to Langflow’s drag-and-drop workflow builder.

​
Prerequisites

  • SambaCloud account and API key.
  • Python 3.10+ (if installing locally). Not required for Langflow Cloud.

​
Setup

1

Install Langflow

Local:
Cloud: Create a free account at langflow.org.Langflow opens at http://localhost:7860 (local) or in your browser (cloud).
2

Create a new flow

Click New Flow and choose a blank canvas or a starter template.
3

Add the SambaNova component

In the component panel, search for SambaNova. Drag the SambaNova text generation component onto the canvas.
4

Configure the component

In the component settings, enter your API key and select a model:
  • SambaNova API Key: your API key from SambaCloud.
  • Model Name: select a model (e.g., Llama-4-Maverick-17B-128E-Instruct).
5

Connect and run

Connect the SambaNova component output to a Chat Output component. Click Run to test your flow.
Langflow canvas with SambaNova component connected to Chat Output

​
Troubleshooting

Verify your API key is correct. Re-enter it in the component settings panel or check it in SambaCloud.
Check supported models and enter the model name manually in the Model Name field.
Ensure langflow run is active, then visit http://localhost:7860. If the port is in use, run langflow run --port 7861.

​
Additional resources