---
title: job_create
url: "/docs/ai-agents/mcp-tools/jobs/job_create/"
canonical_url: https://docs.blossom88.com/docs/ai-agents/mcp-tools/jobs/job_create/
markdown_url: https://docs.blossom88.com/docs/ai-agents/mcp-tools/jobs/job_create.md
summary: Save a Job definition for an Environment.
categories:
- mcp
- reference
---

# job_create

Save a Job definition for an Environment. Saving never runs it; job_run_create does, and an enabled schedule runs it on time.

| | |
|---|---|
| Access | Changes Blossom |
| Same as | [`blossom job:create`](/docs/reference/cli/job/create/) |

## Arguments

| Argument | Type | Description |
|---|---|---|
| `data` | object | The Job, as in job.yml from job:init. Required. |
| `data.environment_id` | string | Environment id. |
| `data.name` | string | — |
| `data.description` | string | — |
| `data.command` | string | Command the Job runs; saving never runs it. |
| `data.source` | object | Optional custom image; the Environment's Release is the default. |
| `data.size` | string | Override size, or inherit to match the Environment's largest Service. |
| `data.resources` | object | Custom size instead of size. |
| `data.timeout_seconds` | integer | Default 900. |
| `data.schedule` | object | Recurring Runs within approved Capacity. |

## Example

```json
{"data":{"clone_url":"https://github.com/acme/shop","services":[{"name":"web","command":"web","port":3000,"publicly_exposed":true}]}}
```