Ctrl K
ComunidadGuías…
GITHUB · ANALIZADO

deckyard

Open-source end-to-end presentation platform that lets you create, present, publish, and collaborate on presentations in a single web-based tool. Self-hosted with full control over your data.

Ficha actualizada 22 sept 2026, 19:24 · upstream 22 sept 2026, 19:20 · release v1.38.1

Ver códigoProblemas y ayudaRepositorio ↗Sitio oficial
00

Qué es deckyard

Descripción editorial construida a partir del repositorio, su documentación y sus señales de mantenimiento.

Deckyard is an open-source, self-hostable presentation platform built with plain Node.js and vanilla ESM. It combines slide creation, live presentation, publishing, and collaboration into one unified experience. Every deck stays yours — from first draft to live audience. Deckyard puts you in control of AI generation by offering full agency over your LLM choices, prompts, and workflow. Built on the Model Context Protocol (MCP), it integrates seamlessly with coding assistants like Claude Desktop or Cursor, allowing agents to build presentations through natural language commands.

CASOS DE USO
  • Create complete presentation decks from raw meeting notes or bullet points
  • Build interactive presentations with live polls and audience follow-along features
  • Present to remote audiences in real-time with built-in Q&A and comment systems
  • Customize presentations for specific audiences using organization-specific themes
  • Embed Deckyard presentations into your own software applications via SDK
  • Create white-label presentations for clients or internal teams without cloud dependencies
PUNTOS DESTACADOS
  • 27 MCP tools covering the full presentation lifecycle: create, iterate, append, convert, compress, analyze, validate, preview, and manage comments
  • 7 guided prompts specifically designed for Claude Desktop that streamline deck creation from structured data to final output
  • Type-aware AI generation pipeline that understands semantic differences between slide types (KPI dashboards vs. timelines) rather than just treating everything as bullet points
  • Full theme customization with organization-specific CSS variables, logo assets, and custom slide type definitions — all stored in a forkable directory structure
PRECAUCIONES
  • Currently in beta; not all features are fully mature or stable across versions
  • No cloud-based hosting option available at this time (requires self-hosted PostgreSQL instance)
  • AI generation quality depends on the underlying LLM provider and prompt engineering
  • Custom slide types require manual configuration of field schemas and render logic in custom directories
01

Perfil de infraestructura

Estimación inicial derivada del repositorio y de los servicios detectados.

CPU4 vCPUrecomendación inicial
MEMORIA8 GBRAM estimada
DISCO30 GBalmacenamiento base
COMPLEJIDADmediumnivel de operación
02

Despliegue detectado

Qué señales reales de instalación hemos encontrado.

Listo para automatizar

Deploy Deckyard using Docker or Node.js. The Docker image includes a local PostgreSQL database, Redis cache, and S3-compatible storage (MinIO/S3). For self-hosting on VPS/VMs, use the provided .env.example file to configure your database connection string, set authentication secrets, and enable MCP server access for agent integration. Start with `npm run setup` to initialize AI keys and authentication, then run `npm run start`. The Docker path requires no additional configuration — Deckyard manages its own database and storage. For remote deployment, use the SSE (Server-Sent Events) transport endpoint at `/mcp` to connect your MCP clients or CI/CD pipelines.

✓Dockerfile
✓Compose
✓One-click
03

Stack y dependencias

Componentes inferidos a partir de manifests, configuración y documentación.

FRAMEWORKS
Node.js
SERVICIOS
PostgreSQLRedisS3/MinIO
BASES DE DATOS
PostgreSQLRedis
PUERTOS
54324177114346379
05

Puede sustituir a

Relaciones detectadas por el análisis del repositorio y su categoría.

06

Experiencia real de la comunidad

Ayuda a contrastar la estimación automática con instalaciones reales.

Todavía no hay datos comunitarios para este proyecto.