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

# LoomLoom overview

> Run repeatable, structured AI workflows from Excel or the command line.

LoomLoom provides an open-source CLI for running repeatable AI workflows through the configured LoomLoom service. It is designed for work that has a consistent input shape, must be applied across multiple rows, and needs traceable results.

The default experience is workbook-first: download a workbook, add one AI Workload per row, validate the file, review the cost estimate, and submit it only after you are ready.

<Columns cols={3}>
  <Card title="Official templates" icon="badge-check" href="/documentation/loomloom/official-templates">
    Start from a CogFoundry-maintained workflow and its generated workbook.
  </Card>

  <Card title="Private templates" icon="lock" href="/documentation/loomloom/private-templates">
    Define and version a workflow for your own recurring process.
  </Card>

  <Card title="SkillBots" icon="bot" href="/documentation/loomloom/skillbots-and-market">
    Discover or publish reviewed workflows through the LoomLoom Market.
  </Card>
</Columns>

## What you can do

* Run official text, image, and video workflows with structured inputs.
* Check the [models currently supported by LoomLoom](/documentation/loomloom/supported-models) for each generation type.
* Create versioned private templates with TemplateSpec.
* Monitor runs and retrieve result workbooks, row results, and generated artifacts.
* Browse, quote, and run Market SkillBots without seeing their private execution logic.
* Install a Market SkillBot or private template version as a local Agent Skill for Codex, Claude Code, or OpenClaw.

## How LoomLoom fits together

```text theme={null}
Template + structured input
          ↓
     hosted run
          ↓
status + result rows + workbook + artifacts
```

The CLI can be used directly. Agent integration is optional and adds guided input collection and confirmation behavior around the same LoomLoom commands.

## Start here

1. [Install LoomLoom](/documentation/loomloom/installation).
2. [Configure your credentials](/documentation/loomloom/configure-credentials).
3. Complete the [Quick Start](/documentation/loomloom/quick-start).
4. Review [Security Notes](/documentation/loomloom/security-notes) before using production data.