{"slug": "promptot-mcp-manage-and-version-llm-prompts-from-your-ai-tools", "title": "PromptOT MCP: Manage and version LLM prompts from your AI tools", "summary": "PromptOT has released an MCP server that allows developers to manage and version LLM prompts directly from AI tools like Claude Desktop, Cursor, and Codex. The server exposes 23 tools for creating, editing, versioning, and testing prompts, enabling prompt operations as part of the development workflow without redeploying applications.", "body_md": "Prompts often start as simple strings in code.\n\nThen the product grows.\n\nYou add a better system prompt. Then a guardrail. Then a different version for production. Then a customer-specific variable. Then test cases. Then someone wants to roll back to the version from last week.\n\nAt that point, prompts are no longer just text. They become part of the application layer.\n\nThat is the problem PromptOT is built around.\n\nPromptOT is a prompt management platform for LLM applications. It lets teams build, version, evaluate, and deliver system prompts through an API without redeploying the app every time a prompt changes.\n\nThe new part I want to share is the **PromptOT MCP server**.\n\nIt lets you manage your PromptOT prompts directly from MCP-compatible AI tools like Claude Desktop, Cursor, Codex, ChatGPT, claude.ai, and other clients.\n\nMCP stands for **Model Context Protocol**.\n\nIt gives AI tools a standard way to connect with external systems. Instead of only chatting with an assistant, you can give the assistant controlled tools for reading and changing real application data.\n\nFor PromptOT, that means your AI assistant can help with things like:\n\nThe goal is not just to generate prompt text. The goal is to manage prompt operations as part of the development workflow.\n\nPromptOT lets you structure prompts as versioned assets instead of hardcoded strings.\n\nA prompt can be composed from typed blocks, such as:\n\nIt also supports variables like:\n\n```\n{{customer_name}}\n{{tone}}\n{{product_context}}\n```\n\nThose variables can be resolved when your application fetches the prompt.\n\nSo instead of redeploying your app every time you adjust a prompt, your app can fetch the latest published prompt through an API.\n\nThe PromptOT MCP server exposes 23 tools across five areas:\n\n| Area | What you can do |\n|---|---|\n| Prompts | List, get, compile, create, update, and delete prompts |\n| Blocks | Create, edit, delete, and reorder typed prompt blocks |\n| Variables | Manage `{{key}}` placeholders |\n| Versions | Save drafts, publish, diff, and roll back |\n| Test cases | Create and manage prompt evaluation inputs |\n\nThat makes it possible to work on prompts from inside the AI tools developers already use.\n\nFor example, you can ask your assistant to:\n\n```\nShow me the current support-agent prompt.\n```\n\nThen:\n\n```\nAdd a stricter refund-policy guardrail as a new block, but save it as a draft.\n```\n\nThen:\n\n```\nCompare this draft with the currently published version.\n```\n\nThen:\n\n```\nPublish the draft if the changes look safe.\n```\n\nThe assistant is not just writing text in a chat window. It is using scoped PromptOT tools to work with real prompt objects.\n\nThe MCP package is available as:\n\n```\nnpx @prompt-ot/mcp\n```\n\nYou will need a scoped PromptOT API key from your PromptOT dashboard.\n\nAdd this to your Claude Desktop MCP config:\n\n```\n{\n  \"mcpServers\": {\n    \"promptot\": {\n      \"command\": \"npx\",\n      \"args\": [\"-y\", \"@prompt-ot/mcp\"],\n      \"env\": {\n        \"PROMPTOT_API_KEY\": \"pot_...\",\n        \"PROMPTOT_MCP_CLIENT\": \"claude-desktop\"\n      }\n    }\n  }\n}\n```\n\nPromptOT also provides a desktop extension bundle for Claude Desktop:\n\n```\nhttps://www.promptot.com/downloads/promptot.mcpb\n```\n\nAdd this to your Cursor MCP config:\n\n```\n{\n  \"mcpServers\": {\n    \"promptot\": {\n      \"command\": \"npx\",\n      \"args\": [\"-y\", \"@prompt-ot/mcp\"],\n      \"env\": {\n        \"PROMPTOT_API_KEY\": \"pot_...\",\n        \"PROMPTOT_MCP_CLIENT\": \"cursor\"\n      }\n    }\n  }\n}\n```\n\nAdd this to your Codex config:\n\n```\n[mcp_servers.promptot]\ncommand = \"npx\"\nargs = [\"-y\", \"@prompt-ot/mcp\"]\nenv = { PROMPTOT_API_KEY = \"pot_...\", PROMPTOT_MCP_CLIENT = \"codex-cli\" }\n```\n\nPromptOT also has a hosted MCP endpoint:\n\n```\nhttps://mcp.promptot.com/mcp\n```\n\nThis can be used by clients that support hosted MCP connections with OAuth.\n\nHere is a realistic workflow.\n\nYou are building an AI support agent. The current system prompt is hardcoded in your app. Every time you want to update the refund policy, escalation rules, or tone, you need to change code, open a pull request, wait for CI, and deploy.\n\nWith PromptOT, you can move that system prompt into a managed prompt.\n\nYou split it into blocks:\n\n```\nRole\nContext\nInstructions\nRefund policy\nEscalation rules\nOutput format\nGuardrails\n```\n\nYour app fetches the published version through the PromptOT API.\n\nThen, with the MCP server connected, you can ask your AI tool to help manage changes:\n\n```\nOpen the support-agent prompt and create a draft version with a stricter escalation rule for billing-related complaints.\n```\n\nThen:\n\n```\nShow me the diff between the published version and this draft.\n```\n\nThen:\n\n```\nCreate a test case for an angry customer asking for a refund outside the refund window.\n```\n\nThen:\n\n```\nPublish the draft.\n```\n\nThis keeps prompt work closer to the tools where developers and AI engineers already think through changes.\n\nPrompt management gets messy because prompts sit between product, engineering, support, and evaluation.\n\nThey are not exactly code, but they affect production behavior like code.\n\nA few things become important quickly:\n\nPromptOT is designed around those questions.\n\nThe MCP server adds another layer: instead of only clicking through a dashboard, you can let your AI development tools interact with the prompt system directly.\n\nPromptOT uses scoped API keys for MCP access.\n\nThat matters because an MCP server can expose powerful actions. PromptOT separates tool capabilities so clients can understand whether a tool is read-only, destructive, or idempotent.\n\nThe goal is to make prompt operations useful without turning every assistant action into an unsafe production change.\n\nLLM apps are moving from simple prompt strings to real prompt infrastructure.\n\nOnce prompts need versions, variables, publishing, rollback, evaluations, and team workflows, they deserve their own operational layer.\n\nPromptOT MCP is one step toward making prompt management feel native inside the AI tools developers already use.", "url": "https://wpnews.pro/news/promptot-mcp-manage-and-version-llm-prompts-from-your-ai-tools", "canonical_source": "https://dev.to/promptot/i-stopped-copy-pasting-prompts-into-chatgpt-heres-the-mcp-setup-that-replaced-it-co", "published_at": "2026-07-18 11:03:46+00:00", "updated_at": "2026-07-18 11:28:24.140309+00:00", "lang": "en", "topics": ["developer-tools", "large-language-models", "ai-tools", "mlops"], "entities": ["PromptOT", "Claude Desktop", "Cursor", "Codex", "ChatGPT", "claude.ai"], "alternates": {"html": "https://wpnews.pro/news/promptot-mcp-manage-and-version-llm-prompts-from-your-ai-tools", "markdown": "https://wpnews.pro/news/promptot-mcp-manage-and-version-llm-prompts-from-your-ai-tools.md", "text": "https://wpnews.pro/news/promptot-mcp-manage-and-version-llm-prompts-from-your-ai-tools.txt", "jsonld": "https://wpnews.pro/news/promptot-mcp-manage-and-version-llm-prompts-from-your-ai-tools.jsonld"}}