The Most Starred Open-Source No-Code AI Tools on GitHub

AutoGPT leads with 180,000 stars, but the no-code AI landscape in 2025 is surprisingly diverse — spanning workflow automation, image generation, local LLM runners, and visual agent builders. The top 15 tools collectively represent over 1.5 million GitHub stars, reflecting explosive growth in accessible AI development platforms.

The most significant finding: image generation tools (Stable Diffusion WebUI at 159k stars and ComfyUI at 97k) rival general-purpose AI platforms, while true no-code solutions like Flowise and AnythingLLM are gaining ground on their low-code counterparts.

Autonomous agents and workflow platforms

AutoGPT — 180,000 stars

**https://github.com/Significant-Gravitas/AutoGPT**

The most-starred AI project on GitHub GitHub has evolved from an experimental autonomous agent into a comprehensive platform. AutoGPT now features a visual block-based workflow builder for designing AI agents, GitHub a marketplace of pre-configured agents, and cloud-hosted deployment options (currently in beta waitlist).

The platform excels at complex multi-step automation — users can build agents that monitor trending topics, generate content, and manage distribution across platforms. github The low-code Agent Builder uses connected blocks where each performs a single action, GitHub making it accessible while retaining power for advanced workflows. Self-hosting requires Docker and technical setup. GitHub

n8n — 156,000 stars

**https://github.com/n8n-io/n8n**

The leading open-source workflow automation platform combines visual building with optional code capabilities. With 400+ integrations and AI-native workflows based on LangChain, n8n bridges traditional automation (Zapier-style) with modern AI agent capabilities. githubGitHub

Key differentiators include the ability to write JavaScript/Python within visual workflows, add npm packages, and self-host with full data control. githubGitHub The platform offers 900+ community templates and enterprise features like SSO and air-gapped deployments. github Classification: low-code, as it explicitly supports custom code alongside the visual interface.

Langflow — 138,000 stars

**https://github.com/langflow-ai/langflow**

A visual platform for building AI-powered agents and workflows with built-in API and MCP (Model Context Protocol) servers. Every workflow can be deployed as a tool integrated into applications on any framework. githubGitHub

The drag-and-drop interface enables rapid prototyping while full Python source code access allows deep customization. Notable features include multi-agent orchestration, integration with observability tools (LangSmith, LangFuse), and a desktop app for Windows/macOS. github Supports all major LLMs including local models. GitHub

Dify — 119,000 stars

**https://github.com/langgenius/dify**

An LLM application development platform combining visual workflow building, RAG pipelines, agent capabilities, and 50+ built-in tools (Google Search, DALL-E, Stable Diffusion, WolframAlpha). github The 1,048-contributor community has made it one of the most actively developed projects.

Dify distinguishes itself with a Prompt IDE for comparing model performance, comprehensive document ingestion (PDF, PPT), and LLMOps features for monitoring and iterating on prompts. github Available as Dify Cloud or self-hosted.

Image generation dominates the no-code space

Stable Diffusion WebUI (AUTOMATIC1111) — 159,000 stars

**https://github.com/AUTOMATIC1111/stable-diffusion-webui**

The definitive interface for Stable Diffusion with a completely visual, truly no-code experience. Users generate images through text prompts, sliders, and buttons — no programming required.

Features include text-to-image, image-to-image, inpainting, outpainting, face restoration, and extensive model support (SD 1.5, SDXL, LoRAs, embeddings). The Gradio-based interface runs locally, supporting NVIDIA, AMD, and Apple Silicon GPUs.

ComfyUI — 97,000 stars

**https://github.com/comfyanonymous/ComfyUI**

A node-based visual workflow builder for image generation that’s more technical than AUTOMATIC1111 but offers greater control. Users create complex pipelines by connecting nodes — no coding, but requires understanding the workflow concept.

Supports Flux, SDXL, and various models with a focus on reproducibility and experimentation. The visual graph approach has made it popular for developing and sharing advanced image generation techniques.

Local LLM runners: Privacy-first AI

Ollama — 156,000 stars

**https://github.com/ollama/ollama**

The Docker-like experience for LLMs: ollama run llama3.3 downloads and runs a model instantly. github Built in Go for performance, Ollama has become the de facto standard for local LLM deployment.

Supports 40+ models (DeepSeek-R1, Gemma 3, Llama 3.3, Mistral, Phi 4), provides an OpenAI-compatible API, and runs on consumer hardware with GPU acceleration. Cross-platform with native apps. github Classification: low-code (CLI-based), but extremely simple.

Open WebUI — 115,000 stars

**https://github.com/open-webui/open-webui**

The most popular web interface for Ollama and OpenAI-compatible APIs. github Features a ChatGPT-like experience that runs entirely offline with built-in RAG, GitHub voice/video calls, web search integration, and multi-user support with RBAC.

Installation is one command via Docker or pip. Supports model building for custom assistants, Python function calling, and image generation integration. Enterprise-ready with SCIM 2.0 provisioning.

GPT4All — 77,000 stars

**https://github.com/nomic-ai/gpt4all**

Nomic AI’s desktop application is truly no-code: download, install, click to download a model, and start chatting. Runs on CPUs (Intel i3 2nd Gen+) without requiring a GPU.

The LocalDocs feature enables private document chat. With 250,000+ monthly active users and the claim of being the “3rd fastest-growing GitHub repo ever,” GPT4All has proven the market for accessible local AI. Python SDK available for developers.

Jan — 40,000 stars

**https://github.com/janhq/jan**

The most polished desktop experience with a beautiful, intuitive interface designed for non-technical users. Native apps for Windows, macOS, and Linux with one-click model downloads from HuggingFace.

Supports both local models (Llama, Gemma, Qwen) and cloud APIs (OpenAI, Anthropic, Mistral, Groq) through a unified interface. Includes MCP integration and a local OpenAI-compatible API server at localhost:1337.

Specialized no-code AI builders

AnythingLLM — 51,000 stars

**https://github.com/Mintplex-Labs/anything-llm**

The “magic box that just works” from Y Combinator-backed Mintplex Labs. A desktop and Docker application with built-in RAG, AI agents, and a no-code agent builder (Agent Flows).

Key features: full MCP compatibility, multi-modal support (images/documents), embeddable chat widgets, support for 30+ LLM providers, and 9 vector database options. github Designed explicitly for non-technical users.

Flowise — 47,000 stars

**https://github.com/FlowiseAI/Flowise**

Purpose-built for no-code LLM application development. The drag-and-drop interface creates chatbots, RAG pipelines, and multi-agent systems in minutes using LangChain under the hood. github

Features ready-to-use templates, conversational memory, and embeddable widgets. Cloud deployment options include AWS, Azure, GCP, Railway, and Render, plus Flowise Cloud for managed hosting. github

Text Generation WebUI (oobabooga) — 45,000 stars

**https://github.com/oobabooga/text-generation-webui**

Called “the definitive Web UI for local AI,” this Gradio interface supports multiple backends (llama.cpp, Transformers, ExLlamaV3, TensorRT-LLM). Portable builds require zero setup — just unzip and run. github

Distinguishes itself with 100% offline operation, zero telemetry, vision/multimodal support, file attachments (PDF, docx), and an OpenAI-compatible API with tool-calling. github

LibreChat — 32,000 stars

**https://github.com/danny-avila/LibreChat**

An enhanced ChatGPT clone combining multiple AI providers (OpenAI, Anthropic, Google, Azure, AWS Bedrock, Ollama) in one self-hosted interface. Features MCP support, code interpreter with sandboxed execution, and a no-code Agent Marketplace. github

Enterprise-ready with OAuth2, LDAP authentication, 30+ language translations, and conversation search/export. github Deployment requires Docker configuration.

Data and infrastructure AI tools

PrivateGPT — 57,000 stars

**https://github.com/zylon-ai/private-gpt**

Production-ready document Q&A with 100% privacy — no data leaves your environment. Built on LlamaIndex with FastAPI, it offers both high-level APIs (document ingestion, chat) and low-level APIs (embeddings, retrieval). GitHub

Supports multiple LLM backends (LlamaCPP, Ollama, OpenAI, Azure, vLLM) and vector databases (Qdrant, Chroma, PGVector). Privategpt Includes a Gradio UI for testing. Classification: low-code, more developer-oriented.

MindsDB — 38,000 stars

**https://github.com/mindsdb/mindsdb**

A federated query engine for AI connecting to 200+ data sources using SQL syntax. Positioned as “the only MCP Server you’ll ever need,” it bridges enterprise databases with AI models. github

Data analysts can query AI using familiar SQL rather than programming languages. GitHub Features built-in Knowledge Bases for unstructured data and scheduling via Jobs. github

LocalAI — 36,000 stars

**https://github.com/mudler/LocalAI**

The OpenAI API drop-in replacement that runs on consumer hardware without GPU requirements. Supports text, image, audio, and video generation with backends including llama.cpp, transformers, vLLM, diffusers, and whisper.

Unique features: P2P distributed inference, AI swarms, MCP support for agents, and a Backend Gallery for installing backends via OCI images. Native apps for all platforms.

How to choose: Decision framework

For pure no-code (zero technical setup):

• GPT4All or Jan for desktop AI chat
• Stable Diffusion WebUI for image generation
• AnythingLLM for all-in-one document/agent capabilities

For visual workflow building:

• Flowise for LLM-specific applications (most accessible)
• ComfyUI for image generation pipelines
• n8n for general automation with AI integration

For developers wanting visual interfaces:

• Langflow for LangChain-based agents
• Dify for comprehensive LLM app development
• AutoGPT for autonomous agent platforms

For enterprise self-hosting:

• Open WebUI for multi-user ChatGPT-like experience
• LibreChat for multi-provider support with authentication
• PrivateGPT for private document intelligence

The distinction between “no-code” and “low-code” matters: tools like Flowise, GPT4All, and Jan require truly zero programming, while n8n, Dify, and Langflow offer visual interfaces but support (and sometimes benefit from) custom code.

Conclusion

The open-source no-code AI ecosystem has matured significantly, with clear category leaders: AutoGPT for autonomous agents, Stable Diffusion WebUI for image generation, n8n/Ollama for infrastructure, and Flowise/AnythingLLM for accessible LLM applications.

The most notable trend is convergence — workflow tools are adding AI capabilities, while AI tools are adding workflow features. MCP (Model Context Protocol) support has become a key differentiator, with Activepieces, AnythingLLM, LibreChat, and LocalAI all highlighting MCP compatibility.

For non-technical users, the best options remain GPT4All (simplest desktop AI), Jan (best interface), and AnythingLLM (most features). For teams building AI applications, Flowise offers the lowest barrier to entry, while Dify and Langflow provide more power for complex use cases.