Self-hosted alternatives to Prompt → Diagram
200 self-hosted apps like Prompt → Diagram, closest match first.
Top picks
- L
LangGraph-GUI
LaunchVisual node-edge graph GUI editor for LangGraph and run with local LLM or online API
FreeOpen sourceSelf-hosted
- S
Sipp
LaunchAI inference, packed simply. A blazing-fast, zero-dependency WebGPU runtime to run GGUF models directly in the browser. Features a symmetric API for seamless local execution and cloud provider routing. Built with Rust & C++.
FreeOpen sourceSelf-hosted
- O

OpenFlowKit
LaunchFree, open-source diagram canvas. Draw it, write it as code (Mermaid, D2, Structurizr in), ask any AI model or a local one, let Claude Code edit it live over MCP, or map a GitHub repo with every arrow cited to a file and line. MIT, local-fi
FreeOpen sourceSelf-hosted
- G

Guaardvark
LaunchSelf-hosted AI studio on your own GPU: chat with your files (RAG), screen and browser agents, coding swarms, LoRA training, an MCP server for Claude Code and Cursor, and local video, image, music and voice generation. No telemetry.
FreeOpen sourceSelf-hosted
- N

Netron
LaunchThis Snap is deprecated. Please migrate to .deb or .rpm. https://github.com/lutzroeder/netron/issues/1500
FreeOpen sourceSelf-hosted
- M
MindGeniusAI
LaunchAn AI agent that reads your PDFs and draws editable mind maps — visible tool-calling loop, built-in RAG, bring-your-own-key, multi-provider (OpenAI / Claude / DeepSeek / Kimi). Self-hostable.
FreeOpen sourceSelf-hosted
More like Prompt → Diagram
Knowledge Graph Extractor
Turn any document or a whole zip into an interactive knowledge graph
About Prompt → Diagram
Prompt → Diagram
Gemma 4 E2B in desktop Chrome (WebGPU)
Prompt to Diagram is a browser demo that turns a text description into an Excalidraw diagram. You describe a system, a flow or a schema, and the Gemma 4 E2B language model writes a short piece of code that is rendered as an editable Excalidraw drawing. The model runs on your own machine through WebGPU, and the page says everything happens in your browser. The demo includes sample prompts such as an OAuth sequence diagram, a CI/CD pipeline, a TCP state machine, an e-commerce schema, payment classes and a checkout swimlane. To reduce memory use, it compresses the model's KV cache about 2.4 times with a TurboQuant-style algorithm rewritten as WGSL compute shaders, which the page says lets longer conversations fit in GPU memory. It needs desktop Chrome 134 or newer with WebGPU subgroups and roughly 3 GB of RAM, and it does not run on Safari or iOS yet. A related turboquant-wasm package, published on npm and GitHub, implements the same algorithm in WASM for vector search on the CPU.
- Category
- AI
- Website
- teamchong.github.io



