We Built the First MCP Server for Cover Letters. Here's Why.
Every job search agent hits the same wall — it can find jobs but falls back to generic LLM calls for the cover letter. We built 5 composable MCP tools that give agents real cover letter intelligence.
TL;DR
There are 177,000+ MCP tools but zero for cover letters. StoryLenses MCP Server gives AI agents 5 composable tools: job analysis, profile matching, narrative-driven letter generation using 15+ archetypes, and quality scoring. Available on the Official MCP Registry, Glama, npm, and as an HTTP endpoint.
The Last Mile Problem
There are now 177,000+ MCP tools across registries. Dice just shipped job search. GitHub, Slack, Notion, databases — they all have MCP servers. You can build an agent that finds jobs, researches companies, and organises applications in a Notion board, all through tool calls.
But the actual application — the document that determines whether a recruiter calls you back — has been left to a raw LLM call and a prayer. The gap is not in the agent's ability to find opportunities. It's in its ability to close them.
What We Built
StoryLenses MCP Server exposes five composable tools that give your agent a complete cover letter workflow:
storylenses_analyze_job — Feed it a job posting URL or text. It extracts 15+ structured fields: the role requirements everyone can see, but also the company's real challenges, culture signals, and what the recruiter actually needs to hear.
storylenses_match_profile — Takes the job analysis and a candidate profile. Returns a match report: fit score, matching skills, gaps, and the strongest narrative angle. This is the step most agents skip entirely and most humans get wrong.
storylenses_generate_letter — Generates a cover letter using the matched data and a narrative archetype selected from 15+ options. The Golden Fleece for the deliberate achiever. The Problem-Solver for someone who thrives in chaos. The Fool Triumphant for the non-traditional candidate.
storylenses_quality_check — Scores the result and returns structured feedback. Because even a good system benefits from a second opinion.
storylenses_list_archetypes — Returns available narrative styles with descriptions, so the agent or user can make an informed choice.
The difference between this and a raw LLM call is the difference between asking a stranger to write your cover letter and asking a career strategist who has read the job posting, studied your CV, and picked the right narrative frame before writing a single word.
Show, Don't Tell
A complete agent workflow takes just four tool calls: analyze the job posting, match the candidate profile, generate the letter with the suggested archetype, and quality-check the result. The analysis, matching, narrative selection, and generation all happen inside StoryLenses. Your agent just orchestrates.
The Reference Agent
We also built an open-source reference agent that chains Dice (job search) with StoryLenses (cover letter generation) into a complete job application workflow. User provides job preferences plus their CV. Agent searches for matching roles. For selected matches, the agent calls StoryLenses to analyze, match, generate, and quality check. Output: a tailored cover letter plus quality score plus match report.
It is designed to be forked. If you are building a career agent, it is your starter kit.
Where to Find It
Available on the Official MCP Registry, Glama, npm (@storylenses/mcp-server), and as an HTTP endpoint. Works with Claude Desktop, Cursor, Windsurf, Claude Code, and any MCP-compatible client.
Learn more at storylenses.app/mcp. Get your API key at storylenses.app/developers. First 10 generations are free.
Ready to try it yourself?
Create a professional, tailored cover letter in minutes.
Write Your First Cover Letter →