What Is Presenton?
Presenton is an open-source AI presentation maker that generates slide decks from prompts, uploaded documents or your own PowerPoint-based designs. The code is on GitHub, where the project describes itself as an alternative to Gamma, Canva and Beautiful AI.
It comes in four forms, per the README: a desktop app, a self-hosted Docker container, the hosted Presenton Cloud in the browser, and a paid Enterprise edition. This review concerns the free desktop and Docker versions.
Get It
Presenton installs as a desktop app or starts from one Docker command; a hosted version also runs in the browser.
Option | How to get it |
|---|---|
| macOS | Download page, .dmg installer |
| Linux | Download page, .AppImage or .deb |
| Windows | Latest .exe is 0.8.7-beta; none in the 0.9.x releases |
| Docker | docker run command in the README, UI at localhost:5001 |
| Website | presenton.ai, also hosts the browser-based Presenton Cloud |
| Privacy policy | presenton.ai/privacy-policy |
| Source code | github.com/presenton/presenton, Apache 2.0 |
Version reviewed: 0.9.11-beta, the latest desktop release, published on 21 September 2026 with .dmg, .AppImage and .deb files. The privacy policy listed covers the Presenton Cloud service, not the desktop or Docker versions.
How to Get Started
The steps follow the README and website; PromptQuorum has not run them.
- 1Install or start it
Why it matters: Open the desktop app, or run docker run -it --name presenton -p 5001:80 -v "./app_data:/app_data" ghcr.io/presenton/presenton:latest and open [localhost:5001](http://localhost:5001) in a browser. - 2Choose your model providers
Why it matters: In the settings, pick a text provider such as Ollama, LM Studio or a cloud API with your own key, and an image provider such as Pexels, Pixabay, ComfyUI or an OpenAI or Gemini image model. - 3Generate and edit
Why it matters: Enter a prompt or upload a document, pick a template, then rearrange and edit the slides in the editor before exporting.
Features Confirmed by the Sources
Every item below comes from the README or the project website; none has been tested by PromptQuorum.
- Inputs and templates. Decks start from a prompt or an uploaded document; built-in templates exist, and a custom template can be generated from an existing PowerPoint file or written in HTML and Tailwind CSS.
- Editing. A drag-and-edit canvas with chat-based slide edits, then export to PPTX or PDF.
- Model providers. Text through Ollama, LM Studio, OpenAI, Google Gemini and Vertex AI, Azure OpenAI, Amazon Bedrock, Anthropic, Fireworks, Together AI or any OpenAI-compatible endpoint, plus a "Sign in with ChatGPT" option that needs no API key.
- Automation. A presentation-generation API and a built-in MCP server, so scripts and AI agents can request decks.
- Teams. Multi-user workspaces with an admin panel in the self-hosted version, and Docker images with GPU support for local models.
Data Flow and Privacy
Where your content goes depends on the providers you configure. The website says the desktop app runs on your machine and that, with local models and local media services, presentation content can stay in your environment, while a cloud text, image or search provider receives the relevant prompts, source content or requests under its own terms.
The README documents two more data paths: a Community gallery of designs that is on by default and contacts a Presenton service (a setting turns it off, and the README requires that for air-gapped use), and an anonymous-telemetry setting that can be disabled. It does not say what the telemetry contains.
📌Note: The privacy policy, last updated on 17 April 2026, covers the hosted Presenton Cloud: it lists account, content and usage data, says personal information is not sold, and says specific presentation content is not used to train models without explicit consent, while aggregated, anonymized usage data may be used. These are the company's statements, not audit results, and PromptQuorum has not inspected the app's network traffic.
Who Should Use It
- Developers and teams who want to generate decks through an API or MCP server on their own infrastructure.
- Privacy-minded presenters who already run Ollama or LM Studio and want slides drafted by a local model.
- People who work in PowerPoint and need an editable .pptx file rather than a web-only deck.
What We Could Not Verify
- Slide quality and speed. No deck was generated, so layout quality, the accuracy of generated text and speed with local models are not assessed.
- Fully offline use. The project describes local operation with Ollama, but PromptQuorum did not confirm that every feature, including image generation and the Community gallery, works without a connection.
- Hardware needs. The sources give no RAM or GPU requirement; it depends on the model you choose.
- Stability. Every release carries a beta label, and how stable updates are between versions is untested.
Competitors and Alternatives
AnythingLLM
- Platforms:
- macOS, Windows, Linux, web
- Price / license:
- Free / MIT
- Key difference:
- Document chat and RAG workspace, not a slide generator
Articles about AnythingLLM (10)
- AnythingLLM Review: Open-Source RAG and Document ChatUpdated September 11, 2026
- Best Local LLM Frontends in 2026: Open WebUI, Enchanted UI, and MoreUpdated October 23, 2026
- AnythingLLM vs PrivateGPT vs Open WebUI: Best Local RAG in 2026Updated October 16, 2026
- Open WebUI in 2026: The Self-Hosted ChatGPT Interface for Ollama and Any AI ModelUpdated October 9, 2026
- Cherry Studio 2.0 in 2026: The Free Agent-Based AI Desktop AppUpdated September 20, 2026
- Local Chat Apps & Assistants Compared (2026): Desktop, Mobile, Self-Hosted, Assistants and RoleplayUpdated September 20, 2026
- Local Knowledge & Retrieval Tools Compared (2026): RAG, Document Chat, Vector Databases, Notes and SearchUpdated September 20, 2026
- Vane Review: Self-Hosted AI Answering Engine (Formerly Perplexica)Updated September 14, 2026
- MaxKB Review: Self-Hosted RAG Platform for DeepSeek and QwenUpdated September 12, 2026
- Onyx Review: Self-Hosted Enterprise Search and RAG (Formerly Danswer)Updated September 12, 2026
+29 more not shown
PrivateGPT
- Platforms:
- macOS, Windows, Linux
- Price / license:
- Free / Apache 2.0
- Key difference:
- Private question answering over your documents; produces answers, not decks
Articles about PrivateGPT (8)
- PrivateGPT Review: It's Not a Chat App AnymoreUpdated September 6, 2026
- AnythingLLM vs PrivateGPT vs Open WebUI: Best Local RAG in 2026Updated October 16, 2026
- Local Knowledge & Retrieval Tools Compared (2026): RAG, Document Chat, Vector Databases, Notes and SearchUpdated September 20, 2026
- Onyx Review: Self-Hosted Enterprise Search and RAG (Formerly Danswer)Updated September 12, 2026
- AnythingLLM Review: Open-Source RAG and Document ChatUpdated September 11, 2026
- AnythingLLM Mobile Review: On-Device Chat + RAGUpdated September 6, 2026
- h2oGPT Review: Features, Setup, and Why the Project Is ArchivedUpdated September 3, 2026
- Best RAG Tools for Business Documents 2026: Local & Private AI ComparedUpdated August 26, 2026
Also mentioned in:
- GPT4All Review: Free Offline AI Chat App by Nomic AIUpdated October 24, 2026
- SurfSense Review: Open-Source Desktop NotebookLM AlternativeUpdated October 21, 2026
+8 more not shown
h2oGPT
- Platforms:
- macOS, Windows, Linux, web
- Price / license:
- Free / Apache 2.0
- Key difference:
- Document chat interface; not built around slide output
Articles about h2oGPT (2)
- h2oGPT Review: Features, Setup, and Why the Project Is ArchivedUpdated September 3, 2026
- Local Knowledge & Retrieval Tools Compared (2026): RAG, Document Chat, Vector Databases, Notes and SearchUpdated September 20, 2026
Also mentioned in:
- Local Inference Engines, Runtimes & Gateways Compared (2026): Run and Serve ModelsUpdated October 24, 2026
No other slide generator is in the PromptQuorum directory yet, so these are the nearest document-driven tools; confirm each app's current price, license and platforms on its own listing.
Frequently Asked Questions
Can I open the exported file in PowerPoint or Google Slides?
The project's FAQ says exported PPTX files are fully editable in Microsoft PowerPoint, Google Slides and compatible software. PromptQuorum did not open an export.
How do I update the desktop app?
Per the download page, installing a newer version replaces the old one and keeps your settings and data.
Does it work with my ChatGPT subscription?
The README says a "Sign in with ChatGPT" option lets a free or paid ChatGPT account generate decks without a separate API key. That route uses OpenAI's service, not a local model.
Can several people share one installation?
Self-hosted installs support multi-user workspaces, each with a private workspace, and an admin panel; per the README, admins manage the shared provider keys.
What does the Enterprise edition add?
The README lists versioned Helm charts for Kubernetes, audit logs, a central admin console and SSO with user provisioning through OAuth, OIDC and SCIM. PromptQuorum did not review it.
Verdict
Presenton is an actively developed, open-source way to generate editable PowerPoint and PDF decks with the AI models you choose, including local ones. Against that, every release is labelled beta, privacy depends entirely on the providers you configure, and nothing about slide quality or speed could be checked here. It suits self-hosters and teams who want control over models and data; anyone who needs dependable output should generate a test deck before relying on it.
Sources
- Presenton website — product overview, download page FAQ and pricing explanation, checked 24 October 2026.
- Presenton on GitHub — README, Apache 2.0 licence and commit history, checked 24 October 2026.
- Presenton releases — file names and dates of the 0.9.11-beta and earlier releases, checked 24 October 2026.
- Presenton privacy policy — data handling of the hosted service, last updated 17 April 2026.
