Self-Hosted AI Apps
The xCloud OneClick catalogue has 53 self-hosted AI apps you can deploy in one click. Pick an app, a domain and a server, and xCloud deploys its containers, volumes, any credentials its template generates and HTTPS; xCloud runs the server and the deployment, and you own the application and its data.
In the catalogue's own words: OpenClaw — Run your personal AI assistant with updates, backups, firewall and more. Hermes Agent — Deploy Hermes Agent in 5 minutes with persistent memory, connected to Telegram, tools, files, APIs and scheduled tasks. Paperclip — Run an autonomous AI company with org charts, budgets and governance. DeepSeek Harness — DeepSeek's own open-source agent harness. Odysseus — Self-hosted chat, agents, memory and research in one private AI workspace, connected to the LLM you choose. Ollama — Run AI models that never leave your server. OpenWebUI — A private AI chat for your whole team. LibreChat — Free open-source AI chat platform with agents, a code interpreter and support for multiple providers. DeepTutor — An open-source AI learning assistant that reads your papers, textbooks and PDFs, then explains, quizzes and reasons over them with a persistent memory of how you learn.
Every tile below links to that app's own page, with what its template sets up and the server plan it is sold on.
All AI apps
Browse all one-click apps →
OpenClawRun your personal AI assistant with updates, backups, firewall and more. No Docker, no hassle — it runs 24/7 on WhatsApp, Telegram, Slack and Discord.
Hermes AgentDeploy Hermes Agent in 5 minutes with persistent memory, connected to Telegram, tools, files, APIs and scheduled tasks. No server setup, no resets.
PaperclipRun an autonomous AI company with org charts, budgets and governance. Hire OpenClaw, Claude Code or Codex agents, live in minutes.
DeepSeek HarnessDeepSeek's own open-source agent harness. Every capability is a plugin you can swap, and every run is an append-only session log you can replay.
OdysseusSelf-hosted chat, agents, memory and research in one private AI workspace, connected to the LLM you choose.
OllamaSelf-hosted runtime for open-source AI models like DeepSeek and Qwen, managed from one place.
OpenWebUICustomizable web interface for AI models, connecting to Ollama and any OpenAI-compatible API endpoint.
LibreChatFree open-source AI chat platform with agents, a code interpreter and support for multiple providers.
DeepTutorAn open-source AI learning assistant that reads your papers, textbooks and PDFs, then explains, quizzes and reasons over them with a persistent memory of how you learn.
OmniRouteAn AI gateway that puts one OpenAI-compatible endpoint in front of hundreds of model providers, with routing strategies, automatic fallback, caching and per-token cost tracking.
Label StudioA data labeling and annotation workspace for images, text, audio, video and structured data, with your training data kept in-house.
MLflowExperiment tracking for machine learning — record runs, compare metrics and parameters, store artifacts and manage early model-lifecycle workflows.
Spice.ai OSSDeploy Spice.ai OSS on xCloud with managed HTTPS, a required API key, federated SQL queries, and verified restart-safe runtime state.
SillyTavernBrowser interface for power users to create character-driven chats with customer-selected AI model providers.
Docling ServeDocument conversion and extraction API. Converts PDFs, Office files, HTML, images, and more into structured Markdown, JSON, or HTML using layout, table, and OCR understanding, with a built-in UI playground.
AnythingLLMAll-in-one, self-hosted AI workspace for chatting with your own documents, building RAG knowledge bases and running AI agents with your own LLM provider.
LibreTranslateFree and open-source machine translation API and web interface, powered by Argos Translate instead of a proprietary cloud service.
EngramSelf-hosted sync server for Engram, giving AI coding agents persistent project memory shared across machines through one HTTPS endpoint.
Crawl4AIOpen-source LLM-friendly web crawler that turns web pages into clean Markdown, with a REST API, browser playground and MCP endpoint.
AnnifOpen-source automated subject indexing for libraries and archives, suggesting subject headings for documents with trained models.
CoaiA self-hosted AI chat gateway with one web UI and an OpenAI-compatible API in front of the model providers you choose.
AstrBotAstrBot v4.28.2 LLM chatbot platform on xCloud with a web dashboard, HTTPS, generated login and persistent data.
Big-AGIMulti-model AI chat workspace on your own domain: bring your own API keys, with chats kept in your browser.
NextChatLight, fast, open-source chat UI for OpenAI, Gemini, Claude, DeepSeek and other model providers.
bolt.diyBrowser-based AI coding workspace — chat with a model provider you choose, then preview and iterate on the app in-browser.
Unstructured APIHTTP API that partitions PDFs, Office files, images and text into structured JSON for search and AI ingestion pipelines.
R2RAn open-source retrieval-augmented generation engine with a REST API, document ingestion, hybrid search, agents and a web dashboard.
OpenVikingAn open-source context database for AI agents that stores hierarchical memory, resources and skills, served over HTTP, MCP and a Studio UI.
LettaAn open-source server for stateful AI agents with persistent memory, exposed as a REST API backed by Postgres with pgvector.
ArgillaAn open-source collaboration tool for AI engineers and domain experts to build and annotate high-quality datasets for NLP, LLM and multimodal projects.
MorphicAn open-source AI answer engine that returns cited, grounded answers through a generative UI, backed by bundled SearXNG search.- VaneA privacy-focused AI answering engine that combines web search with local or cloud LLMs and returns answers with cited sources, with SearXNG bundled.
LobeHubOpen-source AI workspace for chat, agents, knowledge bases and web search, using your own model provider API keys.
Basic MemoryA self-hosted MCP server that gives AI assistants a persistent, searchable Markdown knowledge base with a local knowledge graph.
ScriberrA self-hosted audio and video transcription app producing timestamped transcripts with WhisperX, NVIDIA Parakeet or Canary.
ZeroClawA self-hosted AI agent runtime: a single Rust binary with a web dashboard and gateway, connected to your own Anthropic, OpenAI, Ollama or other OpenAI-compatible provider, reachable from channels such as Discord, Telegram, Matrix, email and webhooks.
FreeLLMAPIAn MIT-licensed gateway that pools the free-tier API keys you already hold from LLM providers behind one OpenAI-compatible endpoint, with automatic routing and failover between providers. It is a proxy over your own keys, not an LLM provider itself.
LocalAISelf-hosted OpenAI-compatible API and web UI that runs language, vision, audio and agent workloads on CPU, with generated API-key auth and persistent model storage.
RAGFlowAn open-source RAG engine that turns your documents into a searchable knowledge base for AI chat and agents, with deep document parsing.
OpenRAGOpen-source retrieval-augmented generation platform: upload documents, then search and chat with them, using OpenSearch, Langflow flows and Docling parsing.
DograhOpen-source voice AI agent platform with a visual workflow builder, call recordings and a TURN relay for browser calls, using your own speech, language and telephony providers.
FirecrawlOpen-source web scraping and crawling API that turns websites into clean Markdown and structured data for LLMs, protected by a generated API key.
ManifestAn open-source, OpenAI-compatible LLM gateway that routes each request across the model providers you connect, with fallbacks, request logs and a dashboard for usage, cost and rate limits.
LlamaGPTSelf-hosted, offline ChatGPT-style chatbot that runs a Llama 2 7B model on your own server CPU, so prompts and answers are not sent to a third party.
faster-whisperA Wyoming speech-to-text server that runs Whisper models on your own server's CPU for Home Assistant voice assistants and other Wyoming clients, with the downloaded model kept on a persistent volume.
SimOpen-source platform for building, deploying and running AI agent workflows on a visual canvas, with knowledge bases, triggers and an API.
HollamaMinimal web chat interface for Ollama and OpenAI-compatible language model servers, with chats kept in each visitor's browser.
SpeakrSelf-hosted audio transcription and note-taking that turns recordings into searchable transcripts, summaries and chat, with persistent upload and instance volumes.
TaskingAIOpen-source platform for building LLM-powered agents and assistants, with a web console, retrieval collections, tools and a REST API on a private PostgreSQL database.
OnyxOpen-source AI chat workspace: connect your own LLM provider, build agents, upload files and organise chats in projects, with private PostgreSQL and HTTPS.
Chatbot UIOpen-source ChatGPT-style web interface for OpenAI and Azure OpenAI models (legacy v1 build), with conversations kept in the browser and no database.
MetaMCPAn MCP aggregator and gateway that combines many MCP servers into namespaces and publishes each as one API-key protected endpoint over SSE, Streamable HTTP or OpenAPI.
LangflowLow-code visual builder for AI agents and RAG workflows: design flows on a canvas, test them in the playground, and call them over the REST API or MCP.
What is self-hosted AI software?
AI software here means applications built around language models and agents: chat front ends, model runners and agent platforms. Self-hosted means you run it on a server you control instead of using it as a service hosted by its vendor: the application, its database and the files it stores sit on that server, so you decide where the data lives, who can reach it and when to move to a new release. The work that usually comes with that, provisioning the server, running the containers and putting HTTPS in front of the app, is what a OneClick deploy takes on.
Frequently Asked Questions
About hosting AI apps on xCloud.
Which server size do I need for a self-hosted AI app?
Each app's page lists the resource floor from its OneClick template and the plan it is sold on. Templates that ask for 2 GB of RAM or less are sold on Cloud VPS 6 GB (6 GB RAM, 4 vCPU cores, 100 GB NVMe SSD); templates that ask for 4 GB or more are sold on Cloud VPS 16 GB (16 GB RAM, 6 vCPU cores, 200 GB NVMe SSD). Several apps can share one server as long as it has the RAM and CPU headroom.
How much does it cost to host AI apps on xCloud?
Cloud VPS 6 GB is $24.99 a month and Cloud VPS 16 GB is $59.99 a month, at a flat monthly price. You pay for the server, so apps that share one server share its price.
Who owns the data in a self-hosted AI app?
You do. The app, its database and its volumes live on the server you chose, under your own account, whether that is an xCloud Cloud VPS or a server you connect yourself. xCloud runs the server and the deployment; the application and its data stay yours, and nothing is routed through a shared multi-tenant service.
Are AI apps upgraded automatically?
No. A OneClick template pins its app image by immutable digest, so new installations use the pinned release and existing sites are not upgraded for you. Back up first, read the upstream release notes, then update the image on your own schedule.
Can I cancel anytime?
Yes. There is no lock-in — cancel any time. Hosting is billed monthly and charged for what you actually use, so when you delete the server the unused balance is refunded to your account, less up to 10% in payment-processing fees.