VLM Run
VLM Run is a visual AI platform that extracts data from images, videos, audio, and documents. It helps automate analysis workflows, such as object detection, transcription, image/audio analysis, and document parsing.
Connect it
In Keystile Automations → Connections → New connection, choose VLM Run and sign in. Your AI tools and workflows then use it through Keystile; they never see the password or key.How connecting works.
Try it from your AI tool
Using Keystile, in VLM Run: analyze audio (Process an audio file, extracting features or transcription).What it can do (6 actions)
| Action | What it does | Changes data |
|---|---|---|
| Analyze Audio | Process an audio file, extracting features or transcription | |
| Analyze Document | Process a document (PDF, DOCX, etc.), extracting structured data or text. | |
| Analyze Image | Process an image (file or URL), extracting descriptions, detecting objects, etc. | |
| Analyze Video | Analyze a video file or URL, e.g. extract frames, detect content, etc. | |
| Custom API Call | Make a custom API call to a specific endpoint | Yes |
| Get File | Gets file's metadata by ID. |
How Keystile protects it
Every VLM Run action is checked before it runs, whether it comes from an AI tool or a workflow. 1 of its actions change data; your rules, limits and budgets decide whether each is allowed, held for a person or stopped, and it all goes into the signed record.
Use VLM Run with your AI assistant — safelyClaude, Cursor or ChatGPT do the work; Keystile checks every action first.
Get early access