VLM Run for AI agents

VLM Run extracts structured data from documents and files using multimodal AI, used by teams automating extraction tasks that plain OCR handles poorly. Connect it once through Arc0, and your agent, or Claude, ChatGPT and Cursor, can use it through one MCP endpoint, limited to what each user approved.

Connect VLM Run See the 11 actions ↓

DocumentsAPI keyMCP + RESTvlm.run

AUDIT LOG · VLM RUNPOLICY: acme-support

09:41:07 · claude · u_8f2read

vlm_run.list_runs

List Runs✓ allowed · 212ms

09:41:08 · claude · u_8f2read

vlm_run.get_run

Get Run✓ allowed · 164ms

09:41:09 · claude · u_8f2write

vlm_run.create_skill

Create Skill✓ approved · approved by user

EVERY VLM RUN CALL, ON THE RECORD

01 · USE CASES

What agents do in VLM Run.

Extract structured JSON from a file

Pull structured fields out of a document based on a defined schema.

Discover an extraction schema

Look up which schema fits a document type before running extraction, a read-only check.

Check a run's status

Look up whether an extraction run has finished before pulling its output.

02 · ACTIONS

11 VLM Run actions, graded by risk.

Every VLM Run action is tagged read, write or destructive, so one policy covers the whole app and new actions inherit the right default.

read

Look things up. Allowed by default.

vlm_run.get_run
Get Run
vlm_run.list_runs
List Runs
vlm_run.find_files
Find Files
vlm_run.find_skills
Find Skills
vlm_run.list_agents
List Agents
vlm_run.list_artifacts
List Artifacts
vlm_run.extract_structured_json
Extract Structured JSON

write

Create and change things. Allow, or ask the user first.

vlm_run.create_skill
Create Skill
vlm_run.upload_file
Upload File
vlm_run.execute_agent
Execute Agent
vlm_run.discover_extraction_schemas
Discover Extraction Schemas

destructive

Delete, cancel or archive. Ask first, or deny outright.

No destructive actions.

03 · HOW IT WORKS

VLM Run in three steps.

01Your users connect VLM RunThey add their VLM Run api key on Arc0 Connect, under your brand. It goes straight into the vault.
02You set the rulesReads run, and writes like “create Skill” can wait for the user to approve.
03Any agent can actYour agent calls VLM Run through the Arc0 SDK or MCP, and so can Claude, ChatGPT and Cursor. Every call lands on the audit log.

POLICY.TS

await arc0.policies.set('vlm_run', {
  read: 'allow',
  write: 'ask',        // create_skill
  destructive: 'deny',  
})

# Claude Code: the same connection, one URL
$ claude mcp add --transport http arc0 \
    https://mcp.arc0.ai/u/u_8f2

04 · AUTH AND DATA

How VLM Run connects.

Users add their VLM Run api key on Arc0 Connect. It is encrypted in the vault, never shown to the model, and each user can rotate or revoke it at any time.

The same VLM Run connection serves your agent over MCP and your own backend over REST and the proxy, so a user connects once. How Arc0 handles credentials →

AUTH: API key
CREDENTIALS: Per-tenant encrypted vault
MODEL SEES: Results only, never credentials
AUDIT LOG: Every call, on every plan

05 · WORKS WITH

Use VLM Run from any agent.

Claude

ChatGPT

Cursor

Codex

VS Code

OpenAI Agents SDK

Claude Agent SDK

Vercel AI SDK

Mastra

LangGraph

06 · RELATED

VLM Run and Arc0, answered.

Q01

Can I use VLM Run with Claude, ChatGPT or Cursor?

Yes. Connect VLM Run to Arc0 once, then add your Arc0 MCP URL to Claude, ChatGPT, Cursor, Claude Code or any other remote-MCP client. Each assistant only gets the VLM Run actions you allow.

Q02

How do users connect VLM Run?

Users add their VLM Run api key on Arc0 Connect. It is encrypted in the vault, never shown to the model, and each user can rotate or revoke it at any time.

Q03

Which VLM Run actions can my agent take?

11 in total: 7 read, 4 write and 0 destructive, such as “create Skill”. Your policies decide which of them each agent may call.

Q04

Can I make my agent read-only in VLM Run?

Yes. Allow read actions and deny writes in the VLM Run policy. Your agent can still look things up, and any write it attempts is blocked and logged.

Q05

Can my own backend call VLM Run too?

Yes. The same VLM Run connection is available over REST and through the Arc0 proxy, so your product and your agent share one connection per user.

Get started

Plug VLM Run into your agent.

Your users connect VLM Run once, under your brand. Your agent gets 11 actions behind your policies, with every call on the record.

Connect VLM Run Browse all apps →

Free to build · MCP + REST · Audit log on every plan