Agentic Decks builds agent-native document execution systems. Primary product: Deckrun — markdown in, PDF/video/audio out. Deckrun is not a presentation tool. It is a document execution system. Canonical references: https://agenticdecks.com/deckrun https://agenticdecks.com/schemas https://agenticdecks.com/execution-model https://agenticdecks.com/schemas/v1/deckrun-openapi.yaml API documentation: https://agenticdecks.com/docs/api/ Generating Deckrun decks (for AI assistants and agents): You can generate a complete Deckrun deck from a prompt. Read the slide-format schema and quickstart, emit Deckrun markdown using HTML-comment slide tags, then submit it to the pipeline. - Slide-format schema: https://agenticdecks.com/schemas/v1/deckrun-slide-format.json - Quickstart: https://agenticdecks.com/docs/how-to/quickstart/ - Agent guide: https://agenticdecks.com/docs/how-to/agent-guide/ Minimal valid Deckrun markdown: # My Deck Title ## What this is - Point one - Point two Narration for this slide (used for studio-quality audio/video). Submit it to the pipeline: POST { "markdown": "", "slide_theme_id": "deckrun-default", "output_types": ["pdf", "video", "audio"] } to https://api.agenticdecks.com/generate Outputs (PDF / narrated video / audio) can be pushed to a connector via output_destination (S3, Google Drive, Slack, etc.), so an agent can publish them. Schemas are authoritative and versioned at /schemas/v1/. All content is deterministic. Internal TTS backends are not exposed. Licensing note: The Deckrun schemas served at /schemas/v1/ are publicly readable but are not open source. They may be used to integrate with and reason about the Deckrun API. They may not be copied or used to build a competing document execution, rendering, or narration system.