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

# AI Integrations

> Use any AI model to generate and edit slides. Gedonus tracks everything regardless of the source.

<span style={{display:'inline-flex',alignItems:'center',padding:'4px 12px',borderRadius:'5px',border:'1px solid oklch(0.62 0.09 62 / 0.38)',background:'oklch(0.78 0.1 62 / 0.18)',color:'oklch(0.46 0.11 62)',fontSize:'10px',fontWeight:'600',letterSpacing:'0.1em',textTransform:'uppercase'}}>In Progress</span>

## How Gedonus works with AI today

Gedonus is **source-agnostic**. It does not care how a change was made — it tracks the result.

If you use Claude for PowerPoint, Microsoft Copilot, a Python script, or a human editor to modify your deck, Gedonus will detect and diff every change when you save the next version. The workflow is the same regardless of the tool.

**Recommended pairing today:**

Use **Claude for PowerPoint** to generate and edit slides, and **Gedonus** to track, diff, and review the output. Claude handles the generation, Gedonus handles the governance.

***

## Supported tools (track changes from any of these)

| Tool                            | Tracking | Notes                            |
| ------------------------------- | -------- | -------------------------------- |
| Claude for PowerPoint           | Yes      | Recommended pairing              |
| Microsoft Copilot               | Yes      | Works inside PowerPoint          |
| GPT-4o (via third-party add-in) | Yes      | Any add-in that edits shapes     |
| Gemini                          | Yes      | Via Google Workspace or add-in   |
| python-pptx scripts             | Yes      | Batch edits tracked on next save |
| Manual edits                    | Yes      | Human edits tracked identically  |

***

## Native AI generation — Coming soon

<span style={{display:'inline-flex',alignItems:'center',padding:'4px 12px',borderRadius:'5px',border:'1px solid oklch(0.73 0.09 93 / 0.48)',background:'oklch(0.958 0.048 93)',color:'oklch(0.37 0.105 93)',fontSize:'10px',fontWeight:'600',letterSpacing:'0.1em',textTransform:'uppercase'}}>Upcoming Release</span>

The planned experience:

1. Type a prompt in the Gedonus task pane (e.g. `Update the revenue chart with Q1 2026 actuals`).
2. Gedonus sends the relevant slide context to your configured AI model.
3. The AI returns a set of shape-level edits.
4. Gedonus shows you the proposed diff before applying anything.
5. You accept, reject, or edit each change shape by shape.
6. Confirmed changes are applied and auto-saved as a new version.

This keeps a human in the loop at every step — no silent overwrite, no mystery changes.

***

## Planned model support

* Claude (Anthropic) — via your own API key
* GPT-4o (OpenAI) — via your own API key
* Gemini (Google) — via your own API key
* Local models (Ollama) — for air-gapped or on-premise environments

Your API key is stored locally. It is never sent to Gedonus servers.

***

## Get notified

<Card title="Request a demo" icon="calendar" href="https://gedonus.com/#demo" horizontal>
  Tell us which AI tools you use and how you use them. It shapes what we build next.
</Card>