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

# Claude skill

> Generate Pixio media from a conversation with Claude

The **Pixio skill** lets Claude create media on your account — ask for an image, a video, or a voice track in conversation and Claude runs the generation through Pixio.

<Card title="pixio-skill on GitHub" icon="github" href="https://github.com/rossman22590/pixio-skill">
  Installation steps, supported commands, and the current version live in the repository README.
</Card>

## What it's good for

* **Generating without leaving your workflow** — you're already in a conversation, you don't switch tabs
* **Letting Claude pick the model** — describe the outcome instead of choosing from 500 models
* **Iterating in language** — "same thing, wider shot, warmer" rather than re-filling a form

## What you need

<Steps>
  <Step title="A Pixio account with credits">
    Generations cost the same credits as they do in the app. Free accounts work — the free balance is enough to try it.
  </Step>

  <Step title="An integration key">
    Create one in **Integrations**. See [API keys](/integrations/api-keys) — give it **Media & workflows** scope at minimum.
  </Step>

  <Step title="The skill installed">
    Follow the README in the repository above. Keep the key in an environment variable, not pasted into a chat.
  </Step>
</Steps>

<Warning>
  Never paste an integration key into a conversation. Configure it where the skill reads it from the environment. Anything typed into a chat is stored in that conversation's history.
</Warning>

## What it can reach

Whatever the key's scopes allow — generation, assets, and saved workflows for a standard key. It's the same account, so:

* Results appear in **My Assets** exactly as if you'd generated them in the app
* Prompts are saved in **Prompt History**
* Costs come out of the same balance
* Your plan's model access applies unchanged

## Limits worth knowing

* **Long video takes minutes.** A conversation isn't the ideal place to wait for a slow render — start it, then check the app.
* **Some surfaces are app-only.** Timeline renders, training jobs, and autonomous Video Agent runs aren't reachable from any integration. See [what you can automate](/integrations/what-you-can-automate).
* **Failed generations are free**, but a run that succeeds and simply isn't what you wanted still costs credits. Iterate on cheap image models first.

<Info>
  The skill is one way in, not a separate product. Anything it makes is a normal Pixio generation you can open, download, chain, or publish from the app.
</Info>