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

# Executions

> Track agent and skill executions end to end on the VLM Run platform

Track agent and skill executions on the VLM Run platform, with details like status, duration, and credit cost. Run single executions or batch workflows, inspect results, and provide feedback to help improve future outputs.

<div className="platform-figure">
  <img src="https://mintcdn.com/autonomiai/31N4NFzIelpYJkPB/images/platform-executions-table.png?fit=max&auto=format&n=31N4NFzIelpYJkPB&q=85&s=51959debeae60919e245123950b519fa" alt="Executions table view" width="2712" height="1512" data-path="images/platform-executions-table.png" />
</div>

## Execution Details

On the left side you can see the request JSON and the inputs you provided to the execution. On the right side you can see the JSON output, along with Markdown and Form views where applicable.

The JSON output lets you quickly verify correctness, and displays bounding box accuracy if grounding was applied to the input.

<div className="platform-figure">
  <img src="https://mintcdn.com/autonomiai/31N4NFzIelpYJkPB/images/platform-executions-27.png?fit=max&auto=format&n=31N4NFzIelpYJkPB&q=85&s=d13891a6c187d6b13afaa09b2e80c18b" alt="Execution detail view" width="1356" height="757" data-path="images/platform-executions-27.png" />
</div>

## JSON Form

The JSON Form tab switches the right panel to a structured view that lets you provide field-level feedback on the execution. For example, if you're analyzing a 10-K and the EPS value is incorrect, you can update it and click save.

<div className="platform-figure">
  <img src="https://mintcdn.com/autonomiai/31N4NFzIelpYJkPB/images/platform-executions-28.png?fit=max&auto=format&n=31N4NFzIelpYJkPB&q=85&s=0b2483cf02cab12c202b90dcfc33518c" alt="Form feedback view" width="1356" height="760" data-path="images/platform-executions-28.png" />
</div>

<Tip>You can also submit field-level feedback through the API. See the [feedback endpoint](/api-reference/v1/post-submit-feedback).</Tip>

## Diff View

Once feedback is provided, you can view a diff by clicking **JSON** and then the diff icon to its left. This makes it easy to scan and understand the changes made through feedback.

<div className="platform-figure">
  <img src="https://mintcdn.com/autonomiai/31N4NFzIelpYJkPB/images/platform-executions-29.png?fit=max&auto=format&n=31N4NFzIelpYJkPB&q=85&s=6c23f189e1102dd06176e7d0dcd39a40" alt="Diff view" width="1356" height="757" data-path="images/platform-executions-29.png" />
</div>

## Execution Feedback

If you don't need field-level feedback, you can provide execution-level feedback by clicking the note button in the top right of the view.

<div className="platform-figure">
  <img src="https://mintcdn.com/autonomiai/31N4NFzIelpYJkPB/images/platform-executions-30.png?fit=max&auto=format&n=31N4NFzIelpYJkPB&q=85&s=d693f100589159d8cb981fba77f45809" alt="Execution feedback" width="1356" height="759" data-path="images/platform-executions-30.png" />
</div>

<Tip>You can also submit execution-level feedback through the API. See the [feedback endpoint](/api-reference/v1/post-submit-feedback).</Tip>

## Executions Table

| Column            | Description                                                            |
| ----------------- | ---------------------------------------------------------------------- |
| **Agent / Skill** | Which agent or skill was executed                                      |
| **Status**        | `running`, `completed`, `failed`, or `cancelled`                       |
| **Duration**      | Total wall-clock time for the entire execution                         |
| **Steps**         | Number of intermediate steps (model calls, tool uses) in the execution |
| **Credits**       | Total credits consumed across all steps                                |
| **Timestamp**     | When the execution was triggered                                       |

Filter by agent, skill, status, or time range. Sort by duration or credits to find outliers.

## Single vs. Batch Executions

The platform supports both:

* **Single executions**: One file, one skill, one result. Ideal for interactive use and testing.
* **Batch executions**: Process multiple files in a single run. The executions table shows the batch as a group, and you can expand it to see individual results.

## Debugging Failures

When an execution fails, the detail view shows:

1. **Which step failed**: Highlighted in the timeline
2. **Error message**: The specific error returned by the model or tool
3. **Input that caused the failure**: The exact payload that triggered the error
4. **Preceding successful steps**: Context for what worked before the failure

This makes it straightforward to isolate whether the issue is in the prompt, the schema, the input file, or a transient infrastructure problem.

## Related Pages

<CardGroup cols={2}>
  <Card title="Observe Overview" icon="chart-line" href="/platform/observe/overview">
    Return to the observability dashboard.
  </Card>

  <Card title="Requests" icon="arrow-right-arrow-left" href="/platform/observe/requests">
    View individual API request logs.
  </Card>

  <Card title="Execute an Agent (API)" icon="robot" href="/api-reference/v1/post-agent-execute">
    Trigger agent executions programmatically.
  </Card>

  <Card title="Artifacts" icon="box-open" href="/agents/artifacts">
    Understand the artifacts produced by executions.
  </Card>
</CardGroup>