Presenton
An AI presentation generator that exports editable PPTX and PDF, with your own models, templates and an API.
Run it in Docker or as a desktop app, bring any model key, and export fully editable PowerPoint.
Presenton generates slide decks from a prompt, a document or your own PowerPoint design, then exports PPTX or PDF. It is for people who want a Gamma-style tool where they pick the model provider and keep the files.
What it is
You describe a deck or upload a document, choose a template, and Presenton writes the slides using the provider you configured, including Ollama, OpenAI, Gemini, Anthropic or any OpenAI-compatible endpoint. Templates are HTML and Tailwind CSS, and a built-in API and MCP server let other tools generate decks.
Who it's for
It suits consultants, educators and teams who want presentations built from their own templates and keys, and developers who want a presentation API. It does not suit anyone who wants zero setup and no model keys; the free route needs Docker or the desktop app plus a provider.
Key features
Bring your own model
Works with Ollama, LM Studio, OpenAI, Gemini, Vertex AI, Azure OpenAI, Bedrock, Anthropic and others.
Editable PPTX export
Exports PowerPoint and PDF, and the README stresses the PPTX stays fully editable.
Templates from PowerPoint
Create custom templates in HTML and Tailwind CSS, or generate one from an existing PowerPoint file.
Presentation API and MCP
Host it as a generation API for your team, with a built-in MCP server.
Multi-user workspaces
Private workspaces per user with an admin panel for accounts and shared keys.
Image options
Choose GPT Image 2, Gemini Flash, Pexels or Pixabay for slide images.
Up and running in 3 steps
Our honest verdict
What's great
- Output is editable PPTX rather than a locked online deck, so you can finish it in PowerPoint.
- You choose and pay the model provider directly, or run Ollama locally for private decks.
- A desktop app for Mac, Windows and Linux is available alongside Docker, so Docker is not mandatory.
- Multi-user admin accounts and an API make it usable as a shared internal service.
Where it falls short
- Self-hosting means running Docker (port 5001) or building the Electron app, and managing your own provider keys and image keys.
- Local models through Ollama need your own hardware, and the README offers GPU Docker variants for that.
- The repository releases are tagged beta (v0.9.x), so expect rough edges.
- Its hosted Presenton Cloud is paid credits-based, and the README notes a Community gallery that makes cloud requests unless you turn it off.
Pricing & license
The self-hosted code is free under Apache-2.0; you pay your own model and image provider fees. Presenton's pricing page shows hosted plans in credits (one slide is one credit): Plus $10/mo for 200 credits, Pro $25/mo for 550, Ultra $50/mo for 1,200 and Max $100/mo for 2,500, with Enterprise by contacting sales (https://presenton.ai/pricing).
Open remote AI-training roles
870 open remote AI-training roles, updated November 14, 2026; most pay $20–$170 an hour. Zekai lists them free; the application goes to the employer’s platform.
Alternatives
These reviewed AI tools do the same job instantly — from our scored directory.
Use it, or have us set it up
FAQ
Can I run it fully offline?
The README says it runs locally and works with Ollama. It also lets you set PRESENTON_COMMUNITY_ENABLED=false to stop the Community gallery making cloud requests.
Which models can generate the slides?
Ollama, LM Studio, OpenAI, Gemini, Vertex AI, Azure OpenAI, Amazon Bedrock, Fireworks, Together AI, Anthropic or any OpenAI-compatible provider.
Is there a hosted version?
Yes, Presenton Cloud runs in the browser and bills in credits; one slide uses one credit. There is also an Enterprise offering with Helm charts, SSO and SCIM.
