# SelfHostMap

> Independent directory of 40 self-hostable open-source AI apps for small and medium businesses (10–200 people). For each app it lists the license and what a business may do with it, the official hardware requirements, an estimated monthly server cost, the official deploy command and known gotchas, each linked to the project's own repo or docs.

- GitHub activity (stars, last commit, latest release) refreshes daily; last refresh Oct 22, 2026.
- Cost estimates = cheapest Hetzner Cloud Regular Performance (shared AMD vCPU) plan meeting the official RAM/vCPU figure, EUR excl. VAT, server only. Apps without an official RAM figure are marked "not estimated", unless we deployed them ourselves; then the estimate is derived from our measured RAM and labelled "based on our test". We never guess.
- License notes are plain-English summaries, not legal advice.
- Full facts for every app in one file: https://selfhostmap.com/llms-full.txt

## Guides (direct answers)

- [Cheapest self-hosted ChatGPT alternatives](https://selfhostmap.com/guides/cheapest-self-hosted-chatgpt-alternatives/): The cheapest self-hosted ChatGPT alternative with an official hardware figure is LibreChat: its minimum 1 vCPU · 1 GB RAM (recommended 2 GB RAM) fits a Hetzner Cloud CPX02 at about €6.49/month (server only).
- [Self-hosted AI apps that run without a GPU](https://selfhostmap.com/guides/self-hosted-ai-apps-without-gpu/): 25 of the 40 apps we track can run on an ordinary CPU server according to their own docs.
- [Self-hosted RAG and knowledge base apps for small businesses](https://selfhostmap.com/guides/best-self-hosted-rag-knowledge-base-apps/): We track 7 self-hostable knowledge-base / RAG apps: RAGFlow, AnythingLLM, PrivateGPT, Open Notebook, Onyx (formerly Danswer), FastGPT, MaxKB.
- [Self-hosted AI apps by license: what a business may do](https://selfhostmap.com/guides/self-hosted-ai-apps-license-commercial-use/): 22 of the 40 apps we track use a permissive license (MIT or Apache-2.0), which allows use, modification and resale with few obligations: Ollama, Langflow, llama.cpp, vLLM, RAGFlow, OpenHands, NextChat, AnythingLLM, PrivateGPT, whisper.cpp, LocalAI, Milvus, LibreChat, Jan, Open Notebook, Vane (formerly Perplexica), Qdrant, Sim, Label Studio, Coze Studio, Paperless-AI, Speaches.

## Tools

- [Self-host vs SaaS AI cost calculator](https://selfhostmap.com/tools/self-host-vs-saas-cost-calculator/): yearly cost of ChatGPT Business, Microsoft 365 Copilot, GitHub Copilot, Notion AI, Zapier and others vs a self-hosted stack (server + model API + setup/upkeep time), break-even months, shareable URL. Every SaaS price links the vendor page and check date; token volumes and labour hours are labelled placeholders.
- [Which self-hosted AI app should we use?](https://selfhostmap.com/tools/which-self-hosted-ai-app/): 5-question wizard (team size, use case, documents, GPU, license) recommending 1-2 catalog apps with data-based reasons.

## Stacks (bundles of catalog apps)

- [Private ChatGPT for a small business (Open WebUI + LiteLLM)](https://selfhostmap.com/stacks/private-chatgpt-for-small-business/): Open WebUI + LiteLLM + Qdrant (optional). Sizing: 1 vCPU · 4 GB RAM (sum of official minimums of LiteLLM), plus whatever Open WebUI needs (no official figure). Server: Hetzner Cloud CPX22 from EUR 19.99/month excl. VAT (floor).
- [Private document Q&A over your company's knowledge (Onyx)](https://selfhostmap.com/stacks/private-document-qa/): Onyx (formerly Danswer) + LiteLLM (optional) + Ollama (optional). Sizing: 4 vCPU · 10 GB RAM · 32 GB disk (sum of official minimums of Onyx (formerly Danswer)). Server: Hetzner Cloud CPX42 from EUR 69.99/month excl. VAT.
- [Self-hosted AI workflow automation (a Zapier replacement with n8n)](https://selfhostmap.com/stacks/ai-workflow-automation/): n8n + Qdrant (optional) + Ollama (optional). Sizing: None of the required components publishes an official CPU/RAM figure, so we do not estimate a server size.. Server: not estimated.
- [Self-hosted meeting transcription (Speaches + Open WebUI)](https://selfhostmap.com/stacks/meeting-transcription/): Speaches + Open WebUI. Sizing: None of the required components publishes an official CPU/RAM figure, so we do not estimate a server size.. Server: not estimated.
- [Self-hosted coding assistant for your dev team (Tabby)](https://selfhostmap.com/stacks/local-coding-assistant/): Tabby + Ollama (optional). Sizing: Needs a GPU server (Tabby); not estimated.. Server: GPU server, not estimated.
- [Self-hosted customer support AI chatbot (Dify)](https://selfhostmap.com/stacks/customer-support-ai-chatbot/): Dify + LiteLLM (optional). Sizing: 2 vCPU · 4 GB RAM (sum of official minimums of Dify). Server: Hetzner Cloud CPX22 from EUR 19.99/month excl. VAT.
- [Fully offline private AI chat (Ollama + Open WebUI)](https://selfhostmap.com/stacks/offline-private-ai/): Ollama + Open WebUI. Sizing: None of the required components publishes an official CPU/RAM figure, so we do not estimate a server size.. Server: not estimated.

## Self-hosted alternatives to SaaS products

- [Self-hosted alternatives to ChatGPT](https://selfhostmap.com/self-hosted-alternative-to/chatgpt/): Open WebUI, LibreChat, LobeHub (LobeChat), AnythingLLM, Jan. ChatGPT Business (Standard seat) costs $25 per user per month on monthly billing, or $20 per user per month on annual billing, according to OpenAI's official pricing information checked on Oct 22, 2026.
- [Self-hosted alternatives to Zapier](https://selfhostmap.com/self-hosted-alternative-to/zapier/): n8n, Activepieces. Zapier Team (2,000 tasks/month, up to 25 users) costs $103.50 per month for the whole plan on monthly billing, or $69 per month for the whole plan on annual billing, according to Zapier's official pricing information checked on Oct 22, 2026.
- [Self-hosted alternatives to Make](https://selfhostmap.com/self-hosted-alternative-to/make/): n8n, Activepieces. Make (Celonis) does not show a price we could verify for Teams; see the official pricing page.
- [Self-hosted alternatives to GitHub Copilot](https://selfhostmap.com/self-hosted-alternative-to/github-copilot/): Tabby. GitHub Copilot Copilot Business costs $19 per user per month on monthly billing, according to GitHub (Microsoft)'s official pricing information checked on Oct 22, 2026.
- [Self-hosted alternatives to Notion AI](https://selfhostmap.com/self-hosted-alternative-to/notion-ai/): Onyx (formerly Danswer), AnythingLLM, Khoj. Notion Business (includes Notion AI) costs $20 per user per month, according to Notion's official pricing information checked on Oct 22, 2026.
- [Self-hosted alternatives to Perplexity](https://selfhostmap.com/self-hosted-alternative-to/perplexity/): Vane (formerly Perplexica), Khoj. Perplexity Enterprise Pro costs $40 per user per month on monthly billing, according to Perplexity AI's official pricing information checked on Oct 22, 2026.
- [Self-hosted alternatives to Otter.ai](https://selfhostmap.com/self-hosted-alternative-to/otter-ai/): Speaches. Otter.ai Business costs $30 per user per month on monthly billing, or $24 per user per month on annual billing, according to Otter.ai's official pricing information checked on Oct 22, 2026.
- [Self-hosted alternatives to Intercom Fin](https://selfhostmap.com/self-hosted-alternative-to/intercom-fin/): Dify, MaxKB, FastGPT, AnythingLLM. Intercom Fin Essential (includes Fin AI Agent) costs $29 per user per month on annual billing, according to Intercom's official pricing information checked on Oct 22, 2026.
- [Self-hosted alternatives to Zendesk AI](https://selfhostmap.com/self-hosted-alternative-to/zendesk-ai/): Dify, MaxKB, FastGPT, AnythingLLM. Zendesk AI Suite Team (includes AI agents) costs $69 per user per month on monthly billing, or $55 per user per month on annual billing, according to Zendesk's official pricing information checked on Oct 22, 2026.
- [Self-hosted alternatives to Midjourney](https://selfhostmap.com/self-hosted-alternative-to/midjourney/): ComfyUI. Midjourney Pro (required for companies with more than $1M gross revenue per year) costs $60 per user per month on monthly billing, or $48 per user per month on annual billing, according to Midjourney's official pricing information checked on Oct 22, 2026.
- [Self-hosted alternatives to DeepL](https://selfhostmap.com/self-hosted-alternative-to/deepl/): LibreTranslate. DeepL Pro Team costs €24.99 per user per month on annual billing, according to DeepL's official pricing information checked on Oct 22, 2026.
- [Self-hosted alternatives to Glean](https://selfhostmap.com/self-hosted-alternative-to/glean/): Onyx (formerly Danswer), Khoj. Glean does not show a price we could verify for Enterprise; see the official pricing page.
- [Self-hosted alternatives to Microsoft 365 Copilot](https://selfhostmap.com/self-hosted-alternative-to/microsoft-365-copilot/): Open WebUI, LibreChat, Onyx (formerly Danswer). Microsoft 365 Copilot Business (add-on) costs $25.20 per user per month on monthly billing, or $21 per user per month on annual billing, according to Microsoft's official pricing information checked on Oct 22, 2026.

## App alternatives (ranked)

- [n8n alternatives](https://selfhostmap.com/alternatives/n8n/): Activepieces, Dify, Langflow, Sim
- [Ollama alternatives](https://selfhostmap.com/alternatives/ollama/): LocalAI, Jan, llama.cpp, vLLM
- [Dify alternatives](https://selfhostmap.com/alternatives/dify/): Langflow, Activepieces, Sim, Coze Studio, n8n, RAGFlow, MaxKB, FastGPT
- [Open WebUI alternatives](https://selfhostmap.com/alternatives/open-webui/): Jan, NextChat, LibreChat, AnythingLLM, LobeHub (LobeChat)
- [Langflow alternatives](https://selfhostmap.com/alternatives/langflow/): Dify, Sim, Coze Studio, n8n, Activepieces
- [llama.cpp alternatives](https://selfhostmap.com/alternatives/llama-cpp/): LocalAI, Jan, Ollama, vLLM
- [vLLM alternatives](https://selfhostmap.com/alternatives/vllm/): Ollama, llama.cpp, LocalAI
- [RAGFlow alternatives](https://selfhostmap.com/alternatives/ragflow/): AnythingLLM, MaxKB, Dify, Open Notebook, FastGPT, Coze Studio, PrivateGPT, Onyx (formerly Danswer)
- [OpenHands alternatives](https://selfhostmap.com/alternatives/openhands/): Tabby
- [NextChat alternatives](https://selfhostmap.com/alternatives/nextchat/): Open WebUI, Jan, LobeHub (LobeChat), LibreChat
- [LobeHub (LobeChat) alternatives](https://selfhostmap.com/alternatives/lobehub/): LibreChat, AnythingLLM, Open WebUI, NextChat, Jan
- [MinerU alternatives](https://selfhostmap.com/alternatives/mineru/): Paperless-AI
- [AnythingLLM alternatives](https://selfhostmap.com/alternatives/anythingllm/): LibreChat, Open WebUI, RAGFlow, PrivateGPT, Open Notebook, Onyx (formerly Danswer), FastGPT, MaxKB, LobeHub (LobeChat)
- [PrivateGPT alternatives](https://selfhostmap.com/alternatives/privategpt/): RAGFlow, AnythingLLM, Open Notebook, Onyx (formerly Danswer), FastGPT, MaxKB
- [Flowise alternatives](https://selfhostmap.com/alternatives/flowise/): Langflow, Dify, Sim, Coze Studio
- [whisper.cpp alternatives](https://selfhostmap.com/alternatives/whisper-cpp/): Speaches
- [LocalAI alternatives](https://selfhostmap.com/alternatives/localai/): Ollama, llama.cpp, vLLM, Jan
- [Milvus alternatives](https://selfhostmap.com/alternatives/milvus/): Qdrant
- [LibreChat alternatives](https://selfhostmap.com/alternatives/librechat/): Open WebUI, NextChat, LobeHub (LobeChat), Jan, AnythingLLM
- [Jan alternatives](https://selfhostmap.com/alternatives/jan/): LobeHub (LobeChat), Open WebUI, NextChat, LibreChat, Ollama, llama.cpp, LocalAI
- [Open Notebook alternatives](https://selfhostmap.com/alternatives/open-notebook/): AnythingLLM, RAGFlow, MaxKB, PrivateGPT, Onyx (formerly Danswer), FastGPT, Khoj
- [Khoj alternatives](https://selfhostmap.com/alternatives/khoj/): Vane (formerly Perplexica), Open Notebook, Onyx (formerly Danswer)
- [Vane (formerly Perplexica) alternatives](https://selfhostmap.com/alternatives/vane/): Khoj
- [Qdrant alternatives](https://selfhostmap.com/alternatives/qdrant/): Milvus
- [Tabby alternatives](https://selfhostmap.com/alternatives/tabby/): OpenHands
- [Onyx (formerly Danswer) alternatives](https://selfhostmap.com/alternatives/onyx/): AnythingLLM, Open Notebook, MaxKB, RAGFlow, PrivateGPT, FastGPT, Khoj
- [FastGPT alternatives](https://selfhostmap.com/alternatives/fastgpt/): AnythingLLM, RAGFlow, PrivateGPT, Open Notebook, Onyx (formerly Danswer), MaxKB, Dify, Coze Studio
- [Sim alternatives](https://selfhostmap.com/alternatives/sim/): Langflow, Dify, Coze Studio, n8n, Activepieces
- [Activepieces alternatives](https://selfhostmap.com/alternatives/activepieces/): Langflow, n8n, Dify, Sim
- [MaxKB alternatives](https://selfhostmap.com/alternatives/maxkb/): AnythingLLM, Open Notebook, Dify, Coze Studio, RAGFlow, PrivateGPT, FastGPT, Onyx (formerly Danswer)
- [Coze Studio alternatives](https://selfhostmap.com/alternatives/coze-studio/): Langflow, Dify, RAGFlow, MaxKB, Sim, FastGPT
- [Paperless-AI alternatives](https://selfhostmap.com/alternatives/paperless-ai/): MinerU
- [Speaches alternatives](https://selfhostmap.com/alternatives/speaches/): whisper.cpp

## Apps

- [n8n](https://selfhostmap.com/apps/n8n/): Workflow automation with 1,500+ integrations and built-in AI agent nodes. License: Sustainable Use License (fair-code) + n8n Enterprise License for .ee files. Hardware: Not published (measured by us: 607 MiB RAM while working). Cost: from €11.99/mo (based on our test, not official).
- [Firecrawl](https://selfhostmap.com/apps/firecrawl/): API that turns websites into clean Markdown or structured data for LLMs. License: AGPL-3.0 (SDKs and some UI components MIT). Hardware: Not published. Cost: Not estimated (no official RAM figure).
- [Ollama](https://selfhostmap.com/apps/ollama/): The simplest way to download and run open-weight LLMs locally behind an API. License: MIT. Hardware: Not published (measured by us: 1.9 GiB RAM while working). Cost: from €19.99/mo (based on our test, not official).
- [Dify](https://selfhostmap.com/apps/dify/): Visual platform for building LLM apps, RAG chatbots, agents and workflows. License: Dify Open Source License (modified Apache-2.0). Hardware: minimum 2 vCPU · 4 GB RAM. Cost: from €19.99/mo.
- [Open WebUI](https://selfhostmap.com/apps/open-webui/): Feature-rich self-hosted chat UI for Ollama and OpenAI-compatible APIs. License: Open WebUI License (BSD-3-Clause style with branding clause). Hardware: Not published (measured by us: 909 MiB RAM while working). Cost: from €11.99/mo (based on our test, not official).
- [Langflow](https://selfhostmap.com/apps/langflow/): Visual builder for AI agents and RAG flows that you can deploy as APIs or MCP servers. License: MIT. Hardware: minimum 1 vCPU · 1.5 GB RAM (recommended 2 vCPU · 4 GB RAM). Cost: from €11.99/mo.
- [ComfyUI](https://selfhostmap.com/apps/comfyui/): Node-based engine for generating images, video, 3D and audio with open models. License: GPL-3.0. Hardware: minimum 8 GB RAM. Cost: GPU server (not estimated).
- [llama.cpp](https://selfhostmap.com/apps/llama-cpp/): Lightweight C/C++ engine that runs quantized open models on CPU or GPU, with an OpenAI-compatible server. License: MIT. Hardware: Not published (measured by us: 489 MiB RAM while working). Cost: from €11.99/mo (based on our test, not official).
- [vLLM](https://selfhostmap.com/apps/vllm/): High-throughput LLM inference server for GPUs with an OpenAI-compatible API. License: Apache-2.0. Hardware: Not published. Cost: GPU server (not estimated).
- [RAGFlow](https://selfhostmap.com/apps/ragflow/): RAG engine with deep document parsing (OCR, layout and tables) and cited answers. License: Apache-2.0. Hardware: recommended 4 vCPU · 16 GB RAM · 50 GB disk (no separate minimum published). Cost: from €69.99/mo.
- [OpenHands](https://selfhostmap.com/apps/openhands/): Self-hosted control center for AI coding agents that write code, review PRs and run automations. License: MIT. Hardware: recommended 2 vCPU · 4 GB RAM (no separate minimum published). Cost: from €19.99/mo.
- [NextChat](https://selfhostmap.com/apps/nextchat/): Lightweight ChatGPT-style web app you can run with one Docker command. License: MIT. Hardware: Not published (measured by us: 57 MiB RAM while working). Cost: from €6.49/mo (based on our test, not official).
- [LobeHub (LobeChat)](https://selfhostmap.com/apps/lobehub/): Polished multi-model chat and agent workspace with plugins and knowledge bases. License: LobeHub Community License (Apache-2.0 with conditions). Hardware: minimum 2 vCPU · 4 GB RAM · 20 GB disk (recommended 4 vCPU · 8 GB RAM · 50 GB disk). Cost: from €19.99/mo.
- [MinerU](https://selfhostmap.com/apps/mineru/): Document parser that converts PDFs, scans and Office files into clean Markdown or JSON for LLMs and RAG. License: MinerU Open Source License (Apache-2.0 with additional terms). Hardware: minimum 2 GB RAM. Cost: from €11.99/mo.
- [AnythingLLM](https://selfhostmap.com/apps/anythingllm/): All-in-one private chat with your documents, agents and an embeddable widget. License: MIT. Hardware: minimum 2 vCPU · 2 GB RAM · 5 GB disk. Cost: from €19.99/mo.
- [LiteLLM](https://selfhostmap.com/apps/litellm/): AI gateway that puts 100+ LLM providers behind one OpenAI-compatible API with keys and budgets. License: MIT (core) + LiteLLM Enterprise License (enterprise directory). Hardware: minimum 1 vCPU · 4 GB RAM. Cost: from €19.99/mo.
- [PrivateGPT](https://selfhostmap.com/apps/privategpt/): API-first private document Q&A that sits on top of your own LLM server. License: Apache-2.0. Hardware: Not published. Cost: Not estimated (no official RAM figure).
- [Flowise](https://selfhostmap.com/apps/flowise/): Drag-and-drop builder for LLM agents and chatflows (archived August 2026). License: Apache-2.0 (enterprise directory under a commercial license). Hardware: Not published. Cost: Not estimated (no official RAM figure).
- [whisper.cpp](https://selfhostmap.com/apps/whisper-cpp/): Fast C/C++ port of OpenAI Whisper speech recognition with a small HTTP transcription server. License: MIT. Hardware: Not published. Cost: Not estimated (no official RAM figure).
- [LocalAI](https://selfhostmap.com/apps/localai/): OpenAI-compatible API for LLMs, vision, voice and images on any hardware. License: MIT. Hardware: Not published. Cost: Not estimated (no official RAM figure).
- [Milvus](https://selfhostmap.com/apps/milvus/): Scalable open-source vector database for semantic search and RAG at large data volumes. License: Apache-2.0. Hardware: minimum 8 GB RAM (recommended 4 vCPU · 16 GB RAM). Cost: from €35.99/mo.
- [LibreChat](https://selfhostmap.com/apps/librechat/): Multi-provider ChatGPT-style interface with agents, file search and user management. License: MIT. Hardware: minimum 1 vCPU · 1 GB RAM (recommended 2 GB RAM). Cost: from €6.49/mo.
- [Jan](https://selfhostmap.com/apps/jan/): Offline desktop ChatGPT alternative that runs models on each employee's computer. License: Apache-2.0. Hardware: minimum 8 GB RAM · 10 GB disk (recommended 16 GB RAM). Cost: No server (desktop app).
- [Open Notebook](https://selfhostmap.com/apps/open-notebook/): Private, self-hosted alternative to Google NotebookLM with your choice of AI models. License: MIT. Hardware: minimum 4 GB RAM. Cost: from €19.99/mo.
- [Khoj](https://selfhostmap.com/apps/khoj/): Personal and team AI "second brain" for your documents and the web. License: AGPL-3.0. Hardware: Not published. Cost: Not estimated (no official RAM figure).
- [PDFMathTranslate](https://selfhostmap.com/apps/pdfmathtranslate/): PDF translator that keeps the original layout, formulas, charts and tables. License: AGPL-3.0. Hardware: Not published. Cost: Not estimated (no official RAM figure).
- [Vane (formerly Perplexica)](https://selfhostmap.com/apps/vane/): Private, Perplexity-style AI answer engine with cited web sources. License: MIT. Hardware: Not published. Cost: Not estimated (no official RAM figure).
- [Langfuse](https://selfhostmap.com/apps/langfuse/): LLM observability with tracing, prompt management, evaluations and cost tracking. License: MIT (core) + Langfuse Enterprise License (ee directories). Hardware: recommended 4 vCPU · 16 GB RAM · 100 GB disk (no separate minimum published). Cost: from €69.99/mo.
- [Qdrant](https://selfhostmap.com/apps/qdrant/): Fast vector database for semantic search and RAG back-ends. License: Apache-2.0. Hardware: Not published (measured by us: 566 MiB RAM while working). Cost: from €11.99/mo (based on our test, not official).
- [Tabby](https://selfhostmap.com/apps/tabby/): Self-hosted AI coding assistant (Copilot alternative) for your team. License: Apache-2.0 (ee directory under a separate license). Hardware: Not published. Cost: GPU server (not estimated).
- [Onyx (formerly Danswer)](https://selfhostmap.com/apps/onyx/): AI chat and enterprise search connected to 50+ workplace apps. License: MIT (Community Edition) + Onyx Enterprise License (ee directories). Hardware: minimum 4 vCPU · 10 GB RAM · 32 GB disk (recommended 8 vCPU · 16 GB RAM · 500 GB disk). Cost: from €69.99/mo.
- [FastGPT](https://selfhostmap.com/apps/fastgpt/): Knowledge-base Q&A and visual AI workflow platform. License: FastGPT Open Source License (Apache-2.0 with conditions). Hardware: minimum 2 vCPU · 4 GB RAM (recommended 2 vCPU · 8 GB RAM). Cost: from €19.99/mo.
- [Sim](https://selfhostmap.com/apps/sim/): Visual builder for AI agent workflows, with knowledge bases, integrations and an API. License: Apache-2.0. Hardware: minimum 2 vCPU · 12 GB RAM · 20 GB disk (recommended 4 vCPU · 16 GB RAM · 50 GB disk). Cost: from €69.99/mo.
- [Label Studio](https://selfhostmap.com/apps/label-studio/): Data labeling tool for text, images, audio and video, used to build and evaluate AI models. License: Apache-2.0. Hardware: minimum 8 GB RAM (recommended 16 GB RAM · 50 GB disk). Cost: from €35.99/mo.
- [Activepieces](https://selfhostmap.com/apps/activepieces/): Open-source Zapier alternative with AI agents, MCP and human-in-the-loop steps. License: MIT (core) + Activepieces Enterprise License (ee directories). Hardware: Not published. Cost: Not estimated (no official RAM figure).
- [MaxKB](https://selfhostmap.com/apps/maxkb/): Enterprise knowledge-base and agent platform with RAG and workflows. License: GPL-3.0. Hardware: minimum 4 vCPU · 8 GB RAM · 100 GB disk. Cost: from €35.99/mo.
- [Coze Studio](https://selfhostmap.com/apps/coze-studio/): Open-source edition of ByteDance's Coze agent-building platform. License: Apache-2.0. Hardware: minimum 2 vCPU · 4 GB RAM. Cost: from €19.99/mo.
- [LibreTranslate](https://selfhostmap.com/apps/libretranslate/): Self-hosted machine translation API with no dependency on Google or DeepL. License: AGPL-3.0. Hardware: Not published. Cost: Not estimated (no official RAM figure).
- [Paperless-AI](https://selfhostmap.com/apps/paperless-ai/): AI auto-tagging and document chat for Paperless-ngx (currently unmaintained). License: MIT. Hardware: Not published. Cost: Not estimated (no official RAM figure).
- [Speaches](https://selfhostmap.com/apps/speaches/): OpenAI-compatible speech-to-text and text-to-speech server (faster-whisper, Kokoro, Piper). License: MIT. Hardware: Not published. Cost: Not estimated (no official RAM figure).

## Data & compliance (telemetry, calls home, SSO, audit log, RBAC)

- [Compliance table, sortable](https://selfhostmap.com/compliance/): per app, from official docs only; "Not documented" = not verifiable, not "absent". Verified so far: Telemetry 18/40, Calls home 8/40, SSO 20/40, Audit log 12/40, RBAC 19/40, Data residency 6/40.
- n8n: telemetry on by default; calls home yes, by default; SSO Paid / enterprise only (OIDC paid · SAML paid · LDAP paid); audit log paid / enterprise only; RBAC paid / enterprise only (https://selfhostmap.com/apps/n8n/#compliance)
- Firecrawl: telemetry off by default (opt-in); calls home not documented; SSO Not documented; audit log not documented; RBAC not documented (https://selfhostmap.com/apps/firecrawl/#compliance)
- Ollama: telemetry not documented; calls home yes, by default; SSO Not documented; audit log not documented; RBAC not documented (https://selfhostmap.com/apps/ollama/#compliance)
- Dify: telemetry on by default; calls home yes, by default; SSO Paid / enterprise only; audit log not documented; RBAC basic free, more in paid plan (https://selfhostmap.com/apps/dify/#compliance)
- Open WebUI: telemetry off by default (opt-in); calls home yes, by default; SSO Free (OIDC free · LDAP free); audit log yes, free; RBAC yes, free (https://selfhostmap.com/apps/open-webui/#compliance)
- Langflow: telemetry on by default; calls home not documented; SSO Free (OIDC free); audit log yes, free; RBAC no (https://selfhostmap.com/apps/langflow/#compliance)
- ComfyUI: telemetry not documented; calls home no (documented); SSO Not documented; audit log not documented; RBAC not documented (https://selfhostmap.com/apps/comfyui/#compliance)
- llama.cpp: telemetry not documented; calls home not documented; SSO Not documented; audit log not documented; RBAC not documented (https://selfhostmap.com/apps/llama-cpp/#compliance)
- vLLM: telemetry on by default; calls home not documented; SSO Not documented; audit log not documented; RBAC not documented (https://selfhostmap.com/apps/vllm/#compliance)
- RAGFlow: telemetry not documented; calls home not documented; SSO Free (OIDC free); audit log not documented; RBAC basic free, more in paid plan (https://selfhostmap.com/apps/ragflow/#compliance)
- OpenHands: telemetry not documented; calls home not documented; SSO Not documented; audit log not documented; RBAC not documented (https://selfhostmap.com/apps/openhands/#compliance)
- NextChat: telemetry not documented; calls home not documented; SSO Not documented; audit log not documented; RBAC not documented (https://selfhostmap.com/apps/nextchat/#compliance)
- LobeHub (LobeChat): telemetry not documented; calls home not documented; SSO Free (OIDC free); audit log not documented; RBAC not documented (https://selfhostmap.com/apps/lobehub/#compliance)
- MinerU: telemetry on by default; calls home not documented; SSO Not documented; audit log not documented; RBAC not documented (https://selfhostmap.com/apps/mineru/#compliance)
- AnythingLLM: telemetry on by default; calls home no (documented); SSO Not documented; audit log not documented; RBAC not documented (https://selfhostmap.com/apps/anythingllm/#compliance)
- LiteLLM: telemetry none (documented); calls home not documented; SSO Free with limits (OIDC free with limits · SAML free with limits); audit log paid / enterprise only; RBAC paid / enterprise only (https://selfhostmap.com/apps/litellm/#compliance)
- PrivateGPT: telemetry not documented; calls home not documented; SSO Paid / enterprise only (LDAP paid); audit log paid / enterprise only; RBAC paid / enterprise only (https://selfhostmap.com/apps/privategpt/#compliance)
- Flowise: telemetry not documented; calls home not documented; SSO Paid / enterprise only (OIDC paid); audit log not documented; RBAC paid / enterprise only (https://selfhostmap.com/apps/flowise/#compliance)
- whisper.cpp: telemetry not documented; calls home not documented; SSO Not documented; audit log not documented; RBAC not documented (https://selfhostmap.com/apps/whisper-cpp/#compliance)
- LocalAI: telemetry not documented; calls home not documented; SSO Free (OIDC free); audit log not documented; RBAC yes, free (https://selfhostmap.com/apps/localai/#compliance)
- Milvus: telemetry not documented; calls home not documented; SSO Not documented; audit log not documented; RBAC yes, free (https://selfhostmap.com/apps/milvus/#compliance)
- LibreChat: telemetry not documented; calls home not documented; SSO Free (OIDC free · SAML free · LDAP free); audit log not documented; RBAC not documented (https://selfhostmap.com/apps/librechat/#compliance)
- Jan: telemetry off by default (opt-in); calls home not documented; SSO Not documented; audit log not documented; RBAC not documented (https://selfhostmap.com/apps/jan/#compliance)
- Open Notebook: telemetry not documented; calls home not documented; SSO Not supported (OIDC not supported · SAML not supported · LDAP not supported); audit log no; RBAC no (https://selfhostmap.com/apps/open-notebook/#compliance)
- Khoj: telemetry on by default; calls home not documented; SSO Not documented; audit log not documented; RBAC not documented (https://selfhostmap.com/apps/khoj/#compliance)
- PDFMathTranslate: telemetry not documented; calls home yes, by default; SSO Not documented; audit log not documented; RBAC not documented (https://selfhostmap.com/apps/pdfmathtranslate/#compliance)
- Vane (formerly Perplexica): telemetry not documented; calls home not documented; SSO Not documented; audit log not documented; RBAC not documented (https://selfhostmap.com/apps/vane/#compliance)
- Langfuse: telemetry on by default; calls home yes, by default; SSO Free (OIDC free); audit log paid / enterprise only; RBAC basic free, more in paid plan (https://selfhostmap.com/apps/langfuse/#compliance)
- Qdrant: telemetry on by default; calls home not documented; SSO Not applicable; audit log yes, free; RBAC yes, free (https://selfhostmap.com/apps/qdrant/#compliance)
- Tabby: telemetry on by default; calls home not documented; SSO Paid / enterprise only (LDAP paid); audit log not documented; RBAC not documented (https://selfhostmap.com/apps/tabby/#compliance)
- Onyx (formerly Danswer): telemetry on by default; calls home not documented; SSO Not documented; audit log not documented; RBAC paid / enterprise only (https://selfhostmap.com/apps/onyx/#compliance)
- FastGPT: telemetry not documented; calls home not documented; SSO Paid / enterprise only; audit log not documented; RBAC paid / enterprise only (https://selfhostmap.com/apps/fastgpt/#compliance)
- Sim: telemetry on by default; calls home not documented; SSO Free (OIDC free · SAML free); audit log yes, free; RBAC yes, free (https://selfhostmap.com/apps/sim/#compliance)
- Label Studio: telemetry on by default; calls home not documented; SSO Paid / enterprise only (OIDC paid · SAML paid · LDAP paid); audit log paid / enterprise only; RBAC no (https://selfhostmap.com/apps/label-studio/#compliance)
- Activepieces: telemetry on by default; calls home not documented; SSO Paid / enterprise only (SAML paid); audit log paid / enterprise only; RBAC paid / enterprise only (https://selfhostmap.com/apps/activepieces/#compliance)
- MaxKB: telemetry not documented; calls home not documented; SSO Paid / enterprise only (OIDC paid · SAML paid · LDAP paid); audit log paid / enterprise only; RBAC paid / enterprise only (https://selfhostmap.com/apps/maxkb/#compliance)
- Coze Studio: telemetry not documented; calls home not documented; SSO Not documented; audit log not documented; RBAC not documented (https://selfhostmap.com/apps/coze-studio/#compliance)
- LibreTranslate: telemetry not documented; calls home not documented; SSO Not documented; audit log not documented; RBAC not documented (https://selfhostmap.com/apps/libretranslate/#compliance)
- Paperless-AI: telemetry not documented; calls home not documented; SSO Not documented; audit log not documented; RBAC not documented (https://selfhostmap.com/apps/paperless-ai/#compliance)
- Speaches: telemetry not documented; calls home not documented; SSO Not documented; audit log not documented; RBAC not documented (https://selfhostmap.com/apps/speaches/#compliance)

## Weekly changes (license changes, archived repos, major releases)

- [Change log](https://selfhostmap.com/changes/): detected daily from GitHub snapshots since 2026-10-22; never edited by hand.
- [What changed in 2026-43 (Oct 19 – Oct 25, 2026)](https://selfhostmap.com/changes/2026-43/): baseline week, no changes yet

## Categories

- [AI app & agent builders](https://selfhostmap.com/categories/agent-builders/): Visual platforms for building chatbots, RAG apps, agents and AI workflows on top of any LLM. (Dify, Langflow, Flowise, Sim, Coze Studio)
- [Knowledge base & RAG](https://selfhostmap.com/categories/knowledge-base/): Upload company documents and let staff or customers ask questions with cited answers. (RAGFlow, AnythingLLM, PrivateGPT, Open Notebook, Onyx (formerly Danswer), FastGPT, MaxKB)
- [Private ChatGPT-style chat](https://selfhostmap.com/categories/chat-ui/): A self-hosted chat interface for your team, connected to cloud APIs or local models. (Open WebUI, NextChat, LobeHub (LobeChat), LibreChat, Jan)
- [Workflow automation with AI](https://selfhostmap.com/categories/automation/): Zapier-style automation that can call LLMs as steps in a business process. (n8n, Activepieces)
- [Local model runtimes](https://selfhostmap.com/categories/llm-runtime/): Run open-weight LLMs on your own hardware and expose an OpenAI-compatible API. (Ollama, llama.cpp, vLLM, LocalAI)
- [LLM gateway & observability](https://selfhostmap.com/categories/llm-ops/): Route, meter, budget and trace LLM calls across providers. (LiteLLM, Langfuse)
- [AI search & research assistants](https://selfhostmap.com/categories/search-research/): Answer engines that search the web or your files and cite sources. (Khoj, Vane (formerly Perplexica))
- [Coding assistants](https://selfhostmap.com/categories/dev-tools/): Self-hosted code completion and chat for engineering teams. (OpenHands, Tabby)
- [Document processing](https://selfhostmap.com/categories/documents/): AI tagging, classification and Q&A for scanned or archived documents. (MinerU, Paperless-AI)
- [Speech & translation](https://selfhostmap.com/categories/speech-translation/): Transcription, text-to-speech and machine translation you can run in-house. (whisper.cpp, PDFMathTranslate, LibreTranslate, Speaches)
- [Image & media generation](https://selfhostmap.com/categories/image-generation/): Generate and edit images, video and audio with open models. (ComfyUI)
- [Data labeling & vector databases](https://selfhostmap.com/categories/data/): Infrastructure for preparing training data and storing embeddings. (Milvus, Qdrant, Label Studio)
- [Web scraping for AI](https://selfhostmap.com/categories/web-data/): Turn websites into clean, LLM-ready Markdown or structured data for RAG, research and agents. (Firecrawl)

## Optional

- [Compare apps side by side](https://selfhostmap.com/compare/): 73 head-to-head pages, e.g. [Activepieces vs Dify](https://selfhostmap.com/compare/activepieces-vs-dify/), [Activepieces vs Langflow](https://selfhostmap.com/compare/activepieces-vs-langflow/), [Activepieces vs n8n](https://selfhostmap.com/compare/activepieces-vs-n8n/), [Activepieces vs Sim](https://selfhostmap.com/compare/activepieces-vs-sim/), [AnythingLLM vs FastGPT](https://selfhostmap.com/compare/anythingllm-vs-fastgpt/)
- [Methodology](https://selfhostmap.com/about/#method): how facts are verified and costs estimated
- [How we test](https://selfhostmap.com/methodology/): our own Docker deployment tests (n8n, Ollama, Open WebUI, llama.cpp, NextChat, Qdrant so far). Numbers marked "measured" are first-hand measurements of RAM, start-up time, image size and CPU tokens/s, never official figures.
- [About](https://selfhostmap.com/about/): independent project run by Ling Feng (solo founder); AI-assisted research verified against official sources; no paid placements
- [Sitemap](https://selfhostmap.com/sitemap.xml)