skyvern_run_task
Pack: skyvern · Connect: https://pipeworx.io/mcp (see Connect below for a single-pack URL)
No MCP client? Call it directly: GET https://gateway.pipeworx.io/v1/tools/skyvern_run_task for the schema, then POST the same URL with its arguments for the data.
Start a Skyvern browser-agent run against a website that publishes no API (search forms, logins, multi-step lookups), described in plain language, and wait briefly for it. Returns the extracted output if the run finishes within wait_seconds (default 20); otherwise returns finished false and a run_id to continue with skyvern_get_run. Runs take seconds to minutes and are billed to your Skyvern account per step. Requires your own Skyvern API key as _apiKey.
Parameters
| Name | Type | Required | Description |
|---|---|---|---|
prompt | string | yes | What the agent should do and return, e.g. “Search for Wikimedia Foundation and return its EIN”. |
url | string | no | Starting page, e.g. the search form. Omit to let Skyvern choose. |
data_extraction_schema | object | no | Optional JSON Schema for the output; strongly recommended so the result is structured. |
max_steps | number | no | Cap on browser steps (billed per step). Default ${DEFAULT_MAX_STEPS}. |
wait_seconds | number | no | How long to wait for the result before returning the run_id (0 to ${MAX_WAIT_S}, default ${MAX_WAIT_S}). 0 starts the run and returns immediately. |
_apiKey | string | no |
Example call
Arguments
{
"prompt": "Search for Wikimedia Foundation and return its EIN",
"url": "https://apps.irs.gov/app/eos/",
"wait_seconds": 20,
"_apiKey": "your-skyvern-api-key"
}
curl
curl -X POST https://gateway.pipeworx.io/skyvern/mcp \
-H 'Content-Type: application/json' \
-d '{"jsonrpc":"2.0","id":1,"method":"tools/call","params":{"name":"skyvern_run_task","arguments":{"prompt":"Search for Wikimedia Foundation and return its EIN","url":"https://apps.irs.gov/app/eos/","wait_seconds":20,"_apiKey":"your-skyvern-api-key"}}}'
TypeScript (@pipeworx/sdk)
import { Pipeworx } from '@pipeworx/sdk';
const pipeworx = new Pipeworx();
const result = await pipeworx.call('skyvern_run_task', {
"prompt": "Search for Wikimedia Foundation and return its EIN",
"url": "https://apps.irs.gov/app/eos/",
"wait_seconds": 20,
"_apiKey": "your-skyvern-api-key"
});
Response shape
| Field | Type | Description |
|---|---|---|
run_id | string | null | |
status | string | null | |
finished | boolean | |
succeeded | boolean | |
output | unknown | |
failure_reason | string | null | |
step_count | number | null | |
created_at | string | null | |
finished_at | string | null | |
recording_url | string | null | |
screenshot_urls | array | |
app_url | string | null | |
warning | string | null | |
next | string | null | |
max_steps | number | null | |
cancel_requested | boolean | |
source | string | null |
Full JSON Schema
{
"type": "object",
"properties": {
"run_id": {
"type": [
"string",
"null"
]
},
"status": {
"type": [
"string",
"null"
]
},
"finished": {
"type": "boolean"
},
"succeeded": {
"type": "boolean"
},
"output": {},
"failure_reason": {
"type": [
"string",
"null"
]
},
"step_count": {
"type": [
"number",
"null"
]
},
"created_at": {
"type": [
"string",
"null"
]
},
"finished_at": {
"type": [
"string",
"null"
]
},
"recording_url": {
"type": [
"string",
"null"
]
},
"screenshot_urls": {
"type": "array",
"items": {
"type": "string"
}
},
"app_url": {
"type": [
"string",
"null"
]
},
"warning": {
"type": [
"string",
"null"
]
},
"next": {
"type": [
"string",
"null"
]
},
"max_steps": {
"type": [
"number",
"null"
]
},
"cancel_requested": {
"type": "boolean"
},
"source": {
"type": [
"string",
"null"
]
}
},
"additionalProperties": true
}
Connect
Add this to your MCP client config — every tool in the catalog, including this one — or use one-click install buttons:
{
"mcpServers": {
"pipeworx": {
"url": "https://pipeworx.io/mcp"
}
}
}
Connect to just the skyvern pack
{
"mcpServers": {
"skyvern": {
"url": "https://gateway.pipeworx.io/skyvern/mcp"
}
}
}
See Getting Started for client-specific install steps.