Clarifyprompt
Turn rough prompts into platform-aware instructions before you burn tokens on weak AI outputs.
Overview
ClarifyPrompt is an MCP server for the Validate phase that optimizes AI prompts for 58+ platforms using your configured LLM API.
What is this MCP server?
- Optimizes prompts for 58+ platforms grouped in 7 categories
- Supports custom platform definitions beyond the built-in catalog
- Uses your own OpenAI-compatible or Anthropic LLM endpoint via env config
- npm package clarifyprompt-mcp v1.1.3 with stdio transport
- MCP tools for optimization—not a static prompt library skill
- 58+ platforms across 7 categories per server description
- npm package clarifyprompt-mcp version 1.1.3
- 3 required or documented env vars: LLM_API_URL, LLM_MODEL, LLM_API_KEY (secret, optional for Ollama)
Community signal: 5 GitHub stars.
What problem does it solve?
Generic prompts fail on platform-specific limits, so your agent outputs look wrong or get rejected.
Who is it for?
Builders shipping on many AI surfaces who already have an LLM API or local Ollama and want MCP-driven prompt refinement.
Skip if: People who need market research, automatic eval scores, or prompt optimization with zero LLM setup.
What do I get? / Deliverables
You get sharper, platform-tuned prompts ready to paste into agents, APIs, or launch copy workflows.
- Platform-tuned prompt text for target surfaces
- Support for custom platform profiles in the optimizer
- Repeatable MCP-accessible optimization from your agent
Recommended MCP Servers
Journey fit
How it compares
LLM-backed prompt optimizer MCP—not a curated skills marketplace or single static SKILL.md.
Common Questions / FAQ
Who is ClarifyPrompt for?
Solo builders and small teams who publish AI features across many tools and want one MCP entry to refine prompts per platform.
When should I use ClarifyPrompt?
While scoping or building AI touchpoints—before you lock prompts for codegen, chat UX, or multi-channel launch content.
How do I add ClarifyPrompt to my agent?
Install clarifyprompt-mcp from npm, set LLM_API_URL, LLM_MODEL, and LLM_API_KEY (if not using local Ollama), add the stdio server to MCP config, and restart your client.