Presentation API

Build presentations
into your product.

Send a brief with your source material. Track the generation run, then retrieve an editable deck to share.

POST /api/public/v1/generation/runs

curl -X POST \
  https://www.encelade.ai/api/public/v1/generation/runs \
  -H "x-api-key: YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "brief": "Q1 performance review. EMEA grew 34% year over year.",
    "pageCount": "auto"
  }'

Start generation with a single request.

How it works

From source material to shareable link.

01
Send the material
POST a brief with your notes, data, and links. Add audience, tone, and other generation options when you need more control.
02
Let generation run
The request starts an asynchronous run. Poll its status or register a webhook for completion and failure events.
03
Open the presentation
Retrieve the completed result as a web link. Review and edit the presentation, then share it with your audience.

Source material

Start with the work
you already have.

Text & documents: Paste meeting notes, research, and reports into the brief — up to 200,000 characters per request.
Spreadsheets: Paste CSV, TSV, or Markdown tables into the brief. Code parses them, computes totals and ranges, and checks every figure on a generated page against your data.
Web sources: Add links to relevant pages and articles, or turn on research to let the run search the web.

Numbers-heavy work

Built for reports
full of numbers.

Exact figures: Code, not the model, parses your tables and computes totals, counts, and ranges. Every figure on a generated page is checked against your data, and a page with a number your data doesn't contain is written again. Charts beside a table are drawn from its cells and redraw when you edit one.
Wide tables: Ask to keep a table and it comes through cell for cell, up to 40 columns and 40 rows. Type shrinks to fit the page down to 14px, then the table scrolls inside the page; on phones, each row becomes its own labeled card. In the editor, tables go up to 50 × 50, take pastes from Excel or Sheets, and support SUM, AVERAGE, MIN, MAX, and ROUND formulas.
PDF & PowerPoint: Export to PDF or PowerPoint (.pptx) from the editor, or start an export job through the API and poll it for the download link. Pages export exactly as they appear in the deck: each PowerPoint slide is a high-resolution image of its page, so its text isn't editable in PowerPoint. In the app, an uploaded .pptx becomes editable text, table, shape, image, and native chart layers.
Edit with AI: Send a plain-language instruction for an existing deck and get the saved changes back, or preview them with a dry run first. It works over the REST API and MCP.

Inside your product

Decks for your users,
from your product.

Included from Pro: The REST API, MCP server, and webhooks are included in the Pro plan and above. No Enterprise contract needed.
Your brand: Set theme to tenant-theme:<uuid> to generate with your workspace brand theme. Create themes in the app or over MCP from your colors, fonts, logo, and backgrounds, and turn off the “Made with Encelade” badge on any paid plan.
Embed anywhere: Every published deck has an embed URL, https://www.encelade.ai/p/<slug>/embed, that you can place in an iframe on any site. Password-protected decks stay locked.
Access per user: Let each of your users connect their own Encelade workspace through OAuth 2.0 with dynamic client registration and PKCE, or run everything from one workspace with scoped API keys. Tag a project with your customer's email by setting its endUserEmail.
Signed webhooks: Receive generation.completed and generation.failed events over HTTPS, signed with HMAC-SHA256 in the X-Webhook-Signature header when you set a secret. Each delivery gets up to three attempts.

Beyond generation

Part of your workflow.
From start to finish.

Manage the presentation lifecycle: List, create, update, and delete projects through the public API.
Choose the right permissions: Use scoped API keys for project access, generation, and webhook management.
Keep your automation informed: Receive webhook events as generation completes or fails, so your workflow can take its next step.

Building with agents? Explore MCP

Your stack