> ## Documentation Index > Fetch the complete documentation index at: https://developer.upsun.com/llms.txt > Use this file to discover all available pages before exploring further. # AI on Upsun > Deploy AI agents, MCP servers, and intelligent applications on Upsun products # AI on Upsun Let Upsun Dispatch™ plan, code, and review your software with governed AI-agent workflows. Use Upsun Cloud to deploy AI agents, MCP servers, RAG applications, and chatbots. Integrate with Claude, GPT, Gemini, and any LLM API. Agents Agentic Automation ## Upsun Dispatch Governed AI-agent workflows that plan, code, review, and optimize your software. Connect a repository, leave a comment, and Upsun Dispatch takes it from there. #### Turn issues into pull requests Comment @upsun-dispatch implement on a GitHub, Jira, or Linear issue and Upsun Dispatch opens or updates a pull or merge request #### Automated code review Upsun Dispatch reviews GitHub pull requests and GitLab merge requests, publishing checks, comments, and summaries #### Full visibility into every run Monitor run status, logs, timelines, and usage from the Upsun Dispatch console [Explore Upsun Dispatch](https://developer.upsun.com/dispatch.md) [Quickstart](https://developer.upsun.com/dispatch/docs/get-started/quickstart.md) ### Everyone works from the same run Captured as it happens. Visible to the whole team, not just whoever kicked it off. ![Upsun Dispatch run timeline showing an approved plan step and a live code step](https://mintcdn.com/upsun-c9761871/Fg69dHDfguHpp19S/images/dispatch/dispatch-issue.png?fit=max\&auto=format\&n=Fg69dHDfguHpp19S\&q=85\&s=8f2144d86158f130e6050beb57968d5e) ### Findings are prioritized and actionable A summary and recommendation, then findings ranked by impact. ![Upsun Cloud performance-agent task report showing a summary, a recommendation, and a ranked list of findings](https://mintcdn.com/upsun-c9761871/oIyhGBlaLmBf3a2Y/images/ai/perf-agent-report.png?fit=max\&auto=format\&n=oIyhGBlaLmBf3a2Y\&q=85\&s=6b23d4d3dc731f73862dae9a7c80b2a0) New: AI Observability ## Upsun Cloud Performance Agent Find performance bottlenecks without writing a profiling script or reading through hours of trace data. Get a report with actionable fixes. #### Correlates every signal you have Profiling data, live traffic, and your source code, all analyzed in one pass #### Nothing to install Bring your own LLM API key and get a report in minutes #### Findings with suggested diffs Each report ranks bottlenecks and proposes a code fix, not just a graph [Get Started with the Performance Agent](https://developer.upsun.com/docs/get-started/ai/upsun-cloud-performance-agent.md) New: AI Observability ## Upsun Cloud Log-Analyzer Agent Analyze and diagnose errors and anomalies in your logs without scanning raw output by hand. Get a prioritized report in minutes. #### Widens the analysis when needed Analyzes beyond your time window and filters, pulling in data like deployments, traffic, and metrics when a finding needs more context #### Nothing to install Bring your own LLM API key and get a report in minutes #### Findings ranked by priority A summary, a top recommendation, and findings ordered by urgency, not raw log severity [Get Started with the Log-Analyzer Agent](https://developer.upsun.com/docs/get-started/ai/upsun-cloud-log-analyzer-agent.md) ### Errors surfaced, not buried in noise A one-sentence summary, a top recommendation, and findings ranked by urgency. ![Log-analyzer agent task report](https://mintcdn.com/upsun-c9761871/_fA8OM8GVhZbwbwv/images/ai/log-analyzer-report.png?fit=max\&auto=format\&n=_fA8OM8GVhZbwbwv\&q=85\&s=78d4b4d6b7b0b614c842ac27028df938) Tools Model Context Protocol ## Upsun MCP Server Give Claude direct access to your Upsun infrastructure. Manage projects, environments, and deployments through natural language conversations. #### Deploy with natural language Ask Claude to deploy your app, create environments, or scale resources #### Manage infrastructure Query logs, check metrics, and troubleshoot issues conversationally #### Works with Claude Desktop Seamless integration with your existing AI assistant workflow [Get Started with MCP](https://developer.upsun.com/docs/get-started/ai/using-the-mcp.md) [All install options](https://github.com/upsun/upsun-mcp-server#installation-instructions) terminal \# Add the Upsun MCP Server \$ claude mcp add --transport http upsun [https://mcp.upsun.com/mcp](https://mcp.upsun.com/mcp) \# Now ask Claude to deployYou: "Deploy my Next.js app to production" ✓ MCP Server connected\ ✓ Found project: my-nextjs-app\ ✓ Creating environment: main\ ✓ Deploying application...\ ✓ Deployment complete! Claude Code CLI \# Deploy with slash commands claude> /upsun deploy --environment=main Building application 'myapp'...\ Provisioning services...\ ✓ Deployed to main \# Check environment status claude> /upsun status Claude Code Integration ## Upsun Skill for Claude Code Deploy directly from your terminal with Claude Code CLI. Manage your entire Upsun infrastructure without leaving your development environment. #### Slash command interface Simple commands like /upsun deploy for instant deployment #### Integrated workflow Code, test, and deploy without context switching #### Full environment control Create, merge, and manage environments from the command line [View on GitHub](https://github.com/upsun/ai) Apps terminal — bash ## Quick Start: Deploy an AI Agent Get started with a simple Python AI agent that integrates with the Anthropic API. .upsun/config.yaml applications:ai-agent:type: "python:3.12"relationships:database: "db:postgresql"hooks:build: |pip install anthropic flaskweb:commands:start: "python app.py" services:db:type: postgresql:16 app.py import anthropicimport os client = anthropic.Anthropic(api\_key=os.environ\["ANTHROPIC\_API\_KEY"]) message = client.messages.create(model="claude-3-5-sonnet-20241022",max\_tokens=1024,messages=\[ \{"role": "user", "content": "Hello, Claude!"} ]) \# Set your API key as an environment variable \$ upsun variable:create --level environment --name ANTHROPIC\_API\_KEY --value "your-key" \# Deploy to Upsun \$ git push upsun main Building application 'ai-agent'... Provisioning PostgreSQL database... ✓ Deployment complete [https://main-abc123.us-5.platformsh.site/](https://main-abc123.us-5.platformsh.site/) [Full AI Agent Tutorial](https://developer.upsun.com/docs/get-started/ai/aiagent.md) Use Cases ## What You Can Build Deploy any AI application that integrates with LLM APIs. Upsun provides the infrastructure, databases, and networking you need. > [card] [AI Agents & Chatbots](https://developer.upsun.com/docs/get-started/ai/aiagent.md) > > Deploy conversational AI powered by Claude, GPT, or Gemini with persistent chat history and context management. > [card] [MCP Servers](https://developer.upsun.com/docs/get-started/ai/deploy-mcp.md) > > Host Model Context Protocol servers that extend Claude Desktop and other AI tools with custom capabilities. > [card] [RAG Applications](https://developer.upsun.com/posts/hands-on/building-internal-ai-agents-for-our-teams-on-upsun) > > Build retrieval-augmented generation apps with vector databases like Chroma and Qdrant for context-aware AI responses. ## Supported Technologies Build with your preferred stack. Any language, any LLM, any database. #### Runtimes Python Node.js PHP Go Ruby Java #### LLM Providers Claude GPT-4 Gemini Azure OpenAI Bedrock #### AI Frameworks LangChain LlamaIndex Haystack Semantic Kernel #### Vector Databases pgvector Chroma Qdrant Redis MongoDB CPU-optimized for AI orchestration. Call external LLM APIs for GPU-intensive inference. Your app handles business logic, data processing, and UX — cost-effective and easy to scale. Learn More ## Resources & Tutorials > [card] [AI Agent Guide](https://developer.upsun.com/docs/get-started/ai/aiagent.md) > > Complete guide to deploying conversational AI agents with persistent memory > [card] [Deploy MCP Servers](https://developer.upsun.com/docs/get-started/ai/deploy-mcp.md) > > Host Model Context Protocol servers for Claude Desktop and other tools > [card] [Building AI Agents Tutorial](https://developer.upsun.com/posts/hands-on/building-internal-ai-agents-for-our-teams-on-upsun) > > Production RAG application with vector search and intelligent context > [card] [Vector Databases](https://developer.upsun.com/docs/add-services) > > Set up Chroma, Qdrant, or pgvector for embeddings storage > [card] [AI Blog Posts](https://developer.upsun.com/posts/articles/ai.md) > > Latest AI tutorials, best practices, and case studies > [card] [Community Forum](https://community.upsun.com) > > Get help and share your AI projects with the community ## Ready to Deploy Your AI Application? Start building AI agents, MCP servers, and intelligent applications on Upsun today. [Start Tutorial →](https://developer.upsun.com/docs/get-started/ai/aiagent.md) [Browse Examples](https://github.com/orgs/upsun/repositories?q=topic%3Aai) This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.