Skip to content

Compare self-hosted AI apps

Choose two or three apps. The table updates instantly, and the URL can be shared.

App 1 App 2 App 3 (optional)

Select at least two apps.

	Activepieces	AnythingLLM	ComfyUI	Coze Studio	Dify	FastGPT	Flowise	Jan	Khoj	Label Studio	Langflow	Langfuse	LibreChat	LibreTranslate	LiteLLM	LobeHub (LobeChat)	LocalAI	MaxKB	n8n	Ollama	Onyx (formerly Danswer)	Open WebUI	Paperless-AI	PrivateGPT	Qdrant	RAGFlow	Speaches	Tabby	Vane (formerly Perplexica)	vLLM
Category	Workflow automation with AI	Knowledge base & RAG	Image & media generation	AI app & agent builders	AI app & agent builders	Knowledge base & RAG	AI app & agent builders	Private ChatGPT-style chat	AI search & research assistants	Data labeling & vector databases	AI app & agent builders	LLM gateway & observability	Private ChatGPT-style chat	Speech & translation	LLM gateway & observability	Private ChatGPT-style chat	Local model runtimes	Knowledge base & RAG	Workflow automation with AI	Local model runtimes	Knowledge base & RAG	Private ChatGPT-style chat	Document processing	Knowledge base & RAG	Data labeling & vector databases	Knowledge base & RAG	Speech & translation	Coding assistants	AI search & research assistants	Local model runtimes
What it is	Open-source Zapier alternative with AI agents, MCP and human-in-the-loop steps.	All-in-one private chat with your documents, agents and an embeddable widget.	Node-based engine for generating images, video, 3D and audio with open models.	Open-source edition of ByteDance's Coze agent-building platform.	Visual platform for building LLM apps, RAG chatbots, agents and workflows.	Knowledge-base Q&A and visual AI workflow platform.	Drag-and-drop builder for LLM agents and chatflows (archived August 2026).	Offline desktop ChatGPT alternative that runs models on each employee's computer.	Personal and team AI "second brain" for your documents and the web.	Data labeling tool for text, images, audio and video, used to build and evaluate AI models.	Visual builder for AI agents and RAG flows that you can deploy as APIs or MCP servers.	LLM observability with tracing, prompt management, evaluations and cost tracking.	Multi-provider ChatGPT-style interface with agents, file search and user management.	Self-hosted machine translation API with no dependency on Google or DeepL.	AI gateway that puts 100+ LLM providers behind one OpenAI-compatible API with keys and budgets.	Polished multi-model chat and agent workspace with plugins and knowledge bases.	OpenAI-compatible API for LLMs, vision, voice and images on any hardware.	Enterprise knowledge-base and agent platform with RAG and workflows.	Workflow automation with 1,500+ integrations and built-in AI agent nodes.	The simplest way to download and run open-weight LLMs locally behind an API.	AI chat and enterprise search connected to 50+ workplace apps.	Feature-rich self-hosted chat UI for Ollama and OpenAI-compatible APIs.	AI auto-tagging and document chat for Paperless-ngx (currently unmaintained).	API-first private document Q&A that sits on top of your own LLM server.	Fast vector database for semantic search and RAG back-ends.	RAG engine with deep document parsing (OCR, layout and tables) and cited answers.	OpenAI-compatible speech-to-text and text-to-speech server (faster-whisper, Kokoro, Piper).	Self-hosted AI coding assistant (Copilot alternative) for your team.	Private, Perplexity-style AI answer engine with cited web sources.	High-throughput LLM inference server for GPUs with an OpenAI-compatible API.
License	MIT (core) + Activepieces Enterprise License (ee directories)	MIT	GPL-3.0	Apache-2.0	Dify Open Source License (modified Apache-2.0)	FastGPT Open Source License (Apache-2.0 with conditions)	Apache-2.0 (enterprise directory under a commercial license)	Apache-2.0	AGPL-3.0	Apache-2.0	MIT	MIT (core) + Langfuse Enterprise License (ee directories)	MIT	AGPL-3.0	MIT (core) + LiteLLM Enterprise License (enterprise directory)	LobeHub Community License (Apache-2.0 with conditions)	MIT	GPL-3.0	Sustainable Use License (fair-code) + n8n Enterprise License for .ee files	MIT	MIT (Community Edition) + Onyx Enterprise License (ee directories)	Open WebUI License (BSD-3-Clause style with branding clause)	MIT	Apache-2.0	Apache-2.0	Apache-2.0	MIT	Apache-2.0 (ee directory under a separate license)	MIT	Apache-2.0
License type	Open core	Permissive	Copyleft (GPL)	Permissive	Custom / conditional	Custom / conditional	Open core	Permissive	Network copyleft (AGPL)	Permissive	Permissive	Open core	Permissive	Network copyleft (AGPL)	Open core	Custom / conditional	Permissive	Copyleft (GPL)	Source-available	Permissive	Open core	Custom / conditional	Permissive	Permissive	Permissive	Permissive	Permissive	Open core	Permissive	Permissive
License conditions	Everything outside packages/ee and packages/server/api/src/app/ee is MIT. Code in those directories needs a commercial license.	Permissive. Internal and commercial use allowed; keep the copyright notice.	Free to use internally and commercially. If you distribute a modified version, you must release its source under GPL-3.0. Each model has its own license; check before commercial use.	Permissive. Internal and commercial use allowed; keep license and notices.	Commercial use is allowed, including as a backend for your own apps. Extra conditions: you may not run a multi-tenant service (one tenant = one workspace) without written permission, and you may not remove or modify the Dify logo or copyright in the frontend.	Commercial use is allowed. You need a commercial license to run a multi-tenant SaaS similar to FastGPT, or to remove or modify the logo or copyright in the console.	Most of the code is Apache-2.0. Code under packages/server/src/enterprise and files with explicit copyright notices are under a separate commercial license.	Permissive. Internal and commercial use allowed; keep license and notices.	Free for internal and commercial use. If you modify Khoj and let users interact with it over a network, you must offer them the modified source code under AGPL-3.0.	Permissive. Internal and commercial use allowed; keep license and notices.	Permissive. Internal and commercial use allowed; keep the copyright notice.	Everything outside ee/, web/src/ee/ and worker/src/ee/ is MIT. Those directories need a commercial license.	Permissive. Internal and commercial use allowed; keep the copyright notice.	Free for internal and commercial use. If you modify it and let users interact with it over a network, you must offer them the modified source. The name and logo are covered by separate trademark guidelines.	Everything outside the "enterprise/" directory is MIT. That directory has its own license.	Commercial use is allowed as a frontend and backend service without modifying the source. Developing and distributing a derivative work requires a commercial license from LobeHub.	Permissive. Internal and commercial use allowed. Models keep their own licenses.	Free to use internally and commercially. If you distribute a modified version, you must release its source under GPL-3.0.	You may use or modify n8n for your own internal business purposes or for non-commercial use. You may not sell it or offer it to others as a paid service. Files with ".ee" in the name need an n8n Enterprise License.	Permissive. Internal and commercial use allowed. Individual models have their own licenses.	The Community Edition is MIT. Code in "ee" directories is under the Onyx Enterprise License.	BSD-style terms plus a branding condition. You may not alter or remove the "Open WebUI" branding unless your deployment has no more than 50 end users in any rolling 30 days, you have written permission, or you hold an enterprise license.	Permissive. Internal and commercial use allowed; keep the copyright notice.	Permissive. Internal and commercial use allowed; keep license and notices.	Permissive. Internal and commercial use allowed; keep license and notices.	Permissive. Internal and commercial use allowed; keep license and notices.	Permissive. Internal and commercial use allowed. Speech models keep their own licenses.	Everything outside the "ee/" directory is Apache-2.0. The ee/ directory has its own license.	Permissive. Internal and commercial use allowed; keep the copyright notice.	Permissive. Internal and commercial use allowed. Models keep their own licenses.
Status	Active	Active	Active	Active	Active	Active	Archived	Active	Active	Active	Active	Active	Active	Active	Active	Active	Active	Active	Active	Active	Active	Active	Stale (no commit in 180+ days)	Active	Active	Active	Stale (no commit in 180+ days)	Active	Active	Active
GitHub stars	25.2k	67.4k	142k	21.7k	159k	29.9k	55.5k	45.0k	37.8k	28.5k	156k	36.2k	46.1k	17.1k	61.8k	83.4k	49.8k	23.0k	208k	184k	32.8k	156k	6.0k	57.6k	35.2k	92.6k	3.7k	33.9k	37.6k	95.1k
Last commit	Oct 22, 2026	Oct 22, 2026	Oct 22, 2026	Jul 29, 2026	Oct 22, 2026	Oct 22, 2026	Aug 13, 2026	Oct 21, 2026	Aug 2, 2026	Oct 22, 2026	Oct 21, 2026	Oct 22, 2026	Oct 22, 2026	Oct 12, 2026	Oct 22, 2026	Oct 22, 2026	Oct 21, 2026	Oct 21, 2026	Oct 22, 2026	Oct 22, 2026	Oct 22, 2026	Oct 16, 2026	Mar 31, 2026	Oct 20, 2026	Oct 21, 2026	Oct 22, 2026	Apr 18, 2026	Jun 30, 2026	Sep 1, 2026	Oct 22, 2026
Latest release	0.92.3	v1.18.0	v0.42.1	v0.5.1	1.17.1	v4.17.2	[email protected]	v0.8.7	2.0.0-beta.28	1.23.2	v1.12.7	v4.67.0	—	v1.9.6	v1.106.2	v2.2.20	v4.11.0	v2.10.8-lts	[email protected]	v0.40.8	v4.9.2	v0.12.3	v3.0.9	v1.0.1	v1.19.2	v1.0.0-rc2	v0.9.0-rc.3	v0.32.0	v1.12.2	v0.32.0
Minimum hardware	Not published	2 vCPU · 2 GB RAM · 5 GB disk	8 GB RAM	2 vCPU · 4 GB RAM	2 vCPU · 4 GB RAM	2 vCPU · 4 GB RAM	Not published	8 GB RAM · 10 GB disk	Not published	8 GB RAM	1 vCPU · 1.5 GB RAM	Not published	1 vCPU · 1 GB RAM	Not published	1 vCPU · 4 GB RAM	2 vCPU · 4 GB RAM · 20 GB disk	Not published	4 vCPU · 8 GB RAM · 100 GB disk	Not published	Not published	4 vCPU · 10 GB RAM · 32 GB disk	Not published	Not published	Not published	Not published	Not published	Not published	Not published	Not published	Not published
Recommended hardware	Not published	Not published	Not published	Not published	Not published	2 vCPU · 8 GB RAM	Not published	16 GB RAM	Not published	16 GB RAM · 50 GB disk	2 vCPU · 4 GB RAM	4 vCPU · 16 GB RAM · 100 GB disk	2 GB RAM	Not published	Not published	4 vCPU · 8 GB RAM · 50 GB disk	Not published	Not published	Not published	Not published	8 vCPU · 16 GB RAM · 500 GB disk	Not published	Not published	Not published	Not published	4 vCPU · 16 GB RAM · 50 GB disk	Not published	Not published	Not published	Not published
GPU	Not required / not stated	Not required / not stated	The README claims it can run even the largest open models on as little as 4 GB VRAM plus 8 GB RAM using weight streaming. NVIDIA, AMD, Intel, Apple Silicon and Ascend are supported.	Not required / not stated	Not required / not stated	Not required / not stated	Not required / not stated	6 GB VRAM minimum for GPU acceleration	Only for offline chat with local models. 16 GB VRAM recommended; an NVIDIA/AMD GPU or an Apple M1+ machine speeds up responses significantly.	Not required / not stated	Not required / not stated	Not required / not stated	Not required / not stated	Optional CUDA 12.4.1 acceleration via docker-compose.cuda.yml.	Not required / not stated	Not required / not stated	Optional. Images exist for NVIDIA CUDA 12/13, AMD ROCm, Intel oneAPI and CPU-only.	Not required / not stated	Not required / not stated	NVIDIA GPUs need compute capability 5.0+ and driver 550+ (570+ for compute capability 5.0–6.2). CPU-only also works.	Not required / not stated	Not required / not stated	Not required / not stated	Not required / not stated	Not required / not stated	Not required / not stated	Optional (CUDA image). A CPU image is available.	NVIDIA GPU recommended. CodeLlama-7B in int8 needs about 8 GB VRAM. int8 needs compute capability 7.0+ (or 6.1), float16 7.0+, bfloat16 8.0+.	Not required / not stated	NVIDIA: compute capability 7.5 or higher (e.g. T4, RTX 20xx, A100, L4, H100, B200).
Est. server cost	Not estimated	from €19.99/mo	GPU server (not estimated)	from €19.99/mo	from €19.99/mo	from €19.99/mo (recommended spec: €35.99/mo)	Not estimated	No server (desktop app)	Not estimated	from €35.99/mo (recommended spec: €69.99/mo)	from €11.99/mo (recommended spec: €19.99/mo)	from €69.99/mo	from €6.49/mo (recommended spec: €11.99/mo)	Not estimated	from €19.99/mo	from €19.99/mo (recommended spec: €35.99/mo)	Not estimated	from €35.99/mo	Not estimated	Not estimated	from €69.99/mo	Not estimated	Not estimated	Not estimated	Not estimated	from €69.99/mo	Not estimated	GPU server (not estimated)	Not estimated	GPU server (not estimated)
Deploy method	Docker (single container, embedded DB)	Docker	Desktop app or manual install	Docker Compose	Docker Compose	Docker Compose (install script)	npm or Docker Compose	Desktop installer	Docker Compose	Docker	Docker	Docker Compose	Docker Compose	Docker (run script) or pip	Docker	Docker Compose (setup script)	Docker	Docker	Docker	Docker (or install script)	Install script (Docker Compose)	Docker	Docker	uv (Python) or Docker	Docker	Docker Compose	Docker	Docker (CUDA)	Docker	pip / uv (or Docker)
Known gotchas	2 listed	2 listed	2 listed	2 listed	2 listed	2 listed	2 listed	2 listed	2 listed	1 listed	2 listed	2 listed	2 listed	2 listed	2 listed	3 listed	1 listed	2 listed	2 listed	2 listed	2 listed	2 listed	2 listed	2 listed	1 listed	2 listed	1 listed	2 listed	2 listed	1 listed
Details	Full Activepieces guide →	Full AnythingLLM guide →	Full ComfyUI guide →	Full Coze Studio guide →	Full Dify guide →	Full FastGPT guide →	Full Flowise guide →	Full Jan guide →	Full Khoj guide →	Full Label Studio guide →	Full Langflow guide →	Full Langfuse guide →	Full LibreChat guide →	Full LibreTranslate guide →	Full LiteLLM guide →	Full LobeHub (LobeChat) guide →	Full LocalAI guide →	Full MaxKB guide →	Full n8n guide →	Full Ollama guide →	Full Onyx (formerly Danswer) guide →	Full Open WebUI guide →	Full Paperless-AI guide →	Full PrivateGPT guide →	Full Qdrant guide →	Full RAGFlow guide →	Full Speaches guide →	Full Tabby guide →	Full Vane (formerly Perplexica) guide →	Full vLLM guide →

Popular comparisons