selfhost.directory

The index of self-hostable projects — ranked, categorized and served with live updates.
Own your data and stop giving it to big tech.

next-ai-draw-io logo

Self-host next-ai-draw-io

v0.5.1

AI chat that draws and edits draw.io diagrams

BetaOpen SourceApache-2.0
docker run -d -p 3000:3000 -e AI_PROVIDER=openai -e AI_MODEL=gpt-4o -e OPENAI_API_KEY=your_api_key ghcr.io/dayuanjiang/next-ai-draw-io:latest
Updated2 weeks ago
LanguageTypescript
Setup difficultyEasy
ThemeDark mode
Stars
36 964
Port
3000 (Web UI; docker run maps it as 3000:3000)
Docker
ghcr.io/dayuanjiang/next-ai-draw-io

What is next-ai-draw-io?

next-ai-draw-io alternatives

Should you self-host next-ai-draw-io?

next-ai-draw-io scores 81/100 on selfhost.directory (how scoring works) and ranks #1 of 12 in Diagramming. It has 36 964 GitHub stars, +788 in the last 30 days.

It shipped 12 stable releases in the past year. The latest, v0.5.1, came out on 31 October 2026. New versions land roughly every 2 weeks.

Setup is easy: most people get it running in one sitting. An official Docker image is available. It is Apache-2.0-licensed and written mostly in Typescript.

Its closest alternative on selfhost.directory is drawdb (80/100, 40 327 stars). next-ai-draw-io scores higher overall. See next-ai-draw-io vs drawdb side by side.

Features

  • Admin PanelEnabled only when ADMIN_PASSWORD is set; read-only on serverless platforms.
  • AI diagram generation from natural language
  • AI editing by chat
  • Animated Connectors
  • Bring-your-own API key
  • Cloud icon libraries
  • Diagram export
  • Image and document to diagram
  • MCP server for AI agents
  • Multiple model providers
  • Offline / intranet operation
  • Per-IP daily usage quotasDisabled unless DYNAMODB_QUOTA_TABLE is set.
  • Reasoning display
  • Selection-scoped AI editing
  • Version history and undo

next-ai-draw-io README

Next AI Draw.io

Draw and edit draw.io diagrams by chatting with an AI.

English | 中文 | 日本語

TrendShift

License: Apache 2.0 Sponsor

Live Demo · Docs · Gallery · Desktop App · MCP Server

promo-web.mp4

Describe a diagram in a sentence and the AI draws it on a real draw.io canvas. Edit it by hand like any draw.io file, or select a few shapes and tell the AI what to change. Every AI change is a version you can compare, restore or undo, and the result exports as .drawio, .png or .svg.

Available as a web app, a desktop app for Windows, macOS and Linux, and an MCP server for AI agents such as Claude Code, Cursor and VS Code.

How to install next-ai-draw-io

Install next-ai-draw-io with Docker

Environment variables
AI_PROVIDER, AI_MODEL, OPENAI_API_KEY, AI_MODELS_CONFIG_PATH
docker run -d -p 3000:3000 \
  -e AI_PROVIDER=openai \
  -e AI_MODEL=gpt-4o \
  -e OPENAI_API_KEY=your_api_key \
  ghcr.io/dayuanjiang/next-ai-draw-io:latest

Or use `--env-file .env` with a copy of env.example; `ai-models.json` can be mounted for server-side models.

How to configure next-ai-draw-io

  • AI_PROVIDERbedrockAI Provider Configuration AI_PROVIDER: Which provider to use Options: bedrock, openai, anthropic, google, vertexai, azure, ollama, openrouter, aihubmix, deepseek, siliconflow, sglang, gateway, edgeone, doubao, modelscope, glm, qwen, qiniu, kimi, minimax, novita, mimo, atlascloud, custom Default: bedrock
  • AI_MODELAI_MODEL: The model ID for your chosen provider (REQUIRED) Tip: For a single-provider quick multi-model setup, list comma-separated model IDs. The first one becomes the default and the rest appear in the model picker. For multiple providers or custom apiKeyEnv/baseUrlEnv, use AI_MODELS_CONFIG / ai-models.json instead. There, apiKeyEnv may also be an array of env var names: each request picks one t
  • AWS_REGIONAWS Bedrock Configuration
  • AWS_ACCESS_KEY_ID
  • AWS_SECRET_ACCESS_KEY

next-ai-draw-io releases & version history

@ Every self-hosted release, weekly

One Monday email with the week's stable releases across the directory.

FAQ

Why do I get "No image provided" after uploading an image?

The chosen model has no vision capability; switch to a vision-capable model (names containing vision or vl) and update to at least v0.4.9.

Why does my self-hosted model only produce thinking text and no diagram?

Small local models often fail at the strict tool-calling format, so use a larger model (32B+ recommended) and enable tool calling in the server, e.g. vLLM with --enable-auto-tool-choice.

How do I deploy on an intranet where embed.diagrams.net is unreachable?

Current images bundle their own copy of draw.io served from /drawio, so the browser no longer needs the external domain; if you still see the error, the image was built with the old build-time default and should be rebuilt.

Why does PDF export do nothing in the web version?

The embedded Draw.io cannot export PDF directly because it depends on an external conversion service that does not work inside an iframe; export a PNG and print that to PDF instead.

More in Diagramming

Rate this project

Click to rate — no account needed.

0 comments

Posting as guest
Loading comments…