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

# Creating Jobs

> How to create and configure jobs

A job combines a file with one or more templates to generate personalized copy for each row.

**Job = File + Templates + Row Range**

## Job Creation Wizard

<Steps>
  <Step title="Step 1: Select File">
    Choose from your uploaded CSV files.

    You'll see:

    * File name
    * Row count
    * Column count
    * Quality score

    <Tip>
      Higher quality scores generally produce better personalization.
    </Tip>
  </Step>

  <Step title="Step 2: Choose Row Range">
    Select which rows to process:

    * **All rows:** Process the entire file
    * **Custom range:** Set start and end rows

    <Tip>
      Start with 10-20 rows to test before processing thousands.
    </Tip>
  </Step>

  <Step title="Step 3: Select Templates">
    Choose one or more templates from **your templates** (not the Template Library).

    **If using multiple templates:**

    * Order matters for chaining!
    * Earlier outputs can be used in later templates
    * Drag to reorder if needed

    <Info>
      **Template order matters!** Templates run top-to-bottom. Use `{{OutputColumnName}}` from earlier templates in later ones.
    </Info>
  </Step>

  <Step title="Step 4: Review & Start">
    Review before starting:

    * Total rows to process
    * Templates selected
    * Estimated cost
    * Usage impact

    Click **Start Job** to begin processing.
  </Step>
</Steps>

## Template Selection

### What Templates Are Available

Only templates from **My Templates** appear in the job wizard:

* ✅ Templates you created
* ✅ Templates you duplicated from the Library
* ❌ System templates (must duplicate first)

### Multiple Templates

You can select multiple templates per job:

| Templates | Usage Per Row | Example                 |
| --------- | ------------- | ----------------------- |
| 1         | 1             | Hook only               |
| 2         | 2             | Hook + CTA              |
| 3         | 3             | Hook + Pain Point + CTA |

<Warning>
  Each template × each row = usage. 3 templates on 1,000 rows = 3,000 usage.
</Warning>

### Template Order for Chaining

If you're using template chaining (where Template B uses output from Template A):

1. Put Template A first in the list
2. Put Template B second
3. Template A's output column becomes available to Template B

## Row Range Selection

### All Rows

Process every row in the file. Best for final runs.

### Custom Range

Set specific start and end rows:

* **Start row:** First row to process (1-based)
* **End row:** Last row to process

**Use cases for custom range:**

* Test on first 20 rows
* Process in batches
* Skip header issues
* Re-run specific sections

## Cost Estimation

Before starting, you'll see:

* **Estimated tokens:** Approximate token usage
* **Estimated cost:** Dollar cost (based on model pricing)
* **Usage impact:** How many rows of your limit this consumes

<Note>
  Estimates are approximations. Actual costs may vary slightly based on output lengths.
</Note>

## Pre-Job Checklist

Before clicking Start:

* [ ] Correct file selected?
* [ ] Row range appropriate? (test small first!)
* [ ] Templates in correct order?
* [ ] Templates tested individually?
* [ ] Enough usage remaining?
* [ ] Cost acceptable?