VLM Run icon

VLM Run

Interact with VLM Run API

Actions6

Overview

This node integrates with the VLM Run API to execute an "agent" operation that processes prompts and associated files. Specifically, the Execute Agent operation allows users to send a prompt along with either a single file or multiple files (mapped by custom keys) to an agent for processing. Upon completion, the agent calls back a specified URL.

Common scenarios include:

  • Automating workflows where an AI agent processes input data described by a prompt and related files.
  • Sending documents, images, or other media to an intelligent agent for analysis or transformation.
  • Handling asynchronous processing by providing a callback URL to receive results once the agent finishes.

Practical example:

  • A user wants to run an AI agent that analyzes a PDF report and an accompanying image. They provide a prompt describing the task, upload both files with custom keys, and specify a callback URL to get notified when processing completes.

Properties

Name Meaning
Prompt The text prompt describing the task for the selected agent to perform.
Multiple Files Boolean flag indicating whether to upload multiple files with custom keys or a single file.
File When uploading a single file, this is the reference to the file data from a previous node.
File Mappings When uploading multiple files, this maps custom keys (identifiers) to URLs of previously uploaded files. Each mapping requires a key and a URL.
Callback URL The URL that the agent will call when processing is complete. This is required.

Output

The node outputs JSON data representing the response from the VLM Run API after executing the agent. The structure depends on the API response but generally includes details about the agent execution status or results.

If multiple items are processed, the output is an array of JSON objects, each corresponding to one input item.

No binary data output is produced by this operation.

Dependencies

  • Requires an API key credential for authenticating with the VLM Run API.
  • The node uses internal helper functions to process files and upload them via presigned URLs before invoking the agent.
  • The callback URL must be accessible externally to receive asynchronous notifications from the agent.

Troubleshooting

  • Error: "At least one file mapping is required when multiple files is enabled"
    Occurs if the user enables multiple files but does not provide any file mappings. Solution: Add at least one file mapping with both key and URL.

  • Error: "Both key and URL are required for each file mapping"
    Happens if any file mapping entry is missing the key or URL. Solution: Ensure all mappings have valid keys and URLs.

  • Error: "Failed to obtain uploaded file URL"
    Indicates that the single file upload did not return a public URL. Solution: Verify the file upload succeeded and the API returned a valid URL.

  • General API errors may occur due to invalid credentials, network issues, or malformed requests. Check the API key configuration and network connectivity.

Links and References

Discussion