# 10ML Perfume Finder Review Free MCP server that gives AI assistants expert fragrance knowledge from the 10ML Fragrance Index, no API key required. ## Editor's Score: 74/100 Ease of use is near-perfect: one URL, no account, listed in Claude's official connectors directory, and a stdio fallback for every other client. Value is the maximum possible because the tool is genuinely free with no paid tier and no hidden limits. Capability scores well within its domain (15 tools, 6,000+ perfumes, 285 notes, multilingual output, a 39/40 benchmark result) but loses points for the hard single-domain ceiling: it cannot be steered outside fragrance at all, which limits its usefulness as a general agent component. Delivery takes a serious hit because the benchmark cited is 10ML's own test with no independent third-party replication found, and no user sentiment or external reviews were surfaced in search, making the accuracy claim unverifiable from outside the vendor. ## Pros - Genuinely zero friction: one URL, no API key, no sign-up, and the server is listed in Claude's official connectors directory so it can be added in two clicks. - Editorial depth that generic product APIs lack: 285 explained notes, full note pyramids, curated editor alternatives, longevity and projection data, and 164 guides in ten languages. - Documented benchmark performance: 39 of 40 correct in a published 40-question test of free perfume MCP servers, with a median response time of 232 ms. ## Cons - Fragrance-only scope is a hard ceiling: the 15 tools cover perfume knowledge exclusively, so any agent that needs to branch into adjacent categories (skincare, fashion, home) hits a wall immediately. - No write tools, no cart integration, and no pricing data beyond 10ML's own decant catalogue, which means it cannot power a full purchase flow even within the fragrance vertical. - The index is curated by 10ML's own editors and reflects their catalogue priorities; coverage of niche or regional houses outside the 249 listed may be thin or absent, and there is no public changelog for when entries are added or updated. ## Pricing Free. - Free: 0 EUR Pricing reflects what we saw at the time of review (2026-10). Confirm current pricing on the official site. Category: AI Agents & Automation (https://aitoolseekers.com/category/ai-agents-automation) Official site: https://10ml.com/perfume-mcp-server Last verified by a human: 2026-10-26 Reviewed by a human at AI Tool Seekers, a hand-reviewed AI tools directory. Every listed tool is verified by a person before publishing. Scores are editorial opinions from a fixed rubric (capability, ease of use, value, delivery on promise) and cannot be bought; there is no pay-to-rank. Directory: https://aitoolseekers.com This review: https://aitoolseekers.com/tools/10ml-perfume-finder Methodology: https://aitoolseekers.com/how-verification-works ## Verdict The go-to choice for anyone who wants a working perfume MCP server in under a minute: one URL, no account, and a documented 39/40 benchmark score. Its depth is deliberately narrow, covering fragrance knowledge only, so any agent needing broader product or e-commerce data will outgrow it fast. Developers building general-purpose shopping assistants should look elsewhere; this is a specialist tool, not a platform. Best for: Developers and fragrance enthusiasts adding perfume knowledge to AI assistants Not for: Teams needing broad e-commerce or multi-category product data in their agents ## Overview 10ML Perfume Finder is a Model Context Protocol server that connects AI assistants to the 10ML Fragrance Index, a curated database of 6,000+ perfumes across 249 houses. The single endpoint at https://mcp.10ml.com/mcp requires no authentication and no account. You paste the URL into Claude, ChatGPT, Cursor, or any MCP-compatible client and the assistant immediately gains 15 read-only fragrance tools. Setup is measured in seconds, not minutes. What separates it from a generic product-search API is editorial depth. The index includes full note pyramids (top, heart, base), accords, family classifications, longevity and projection data, 285 explained notes, 164 guides in ten languages, and curated editor alternatives to famous perfumes. That is not scraped data; 10ML's fragrance editors maintain it. The server scored 39 of 40 correct in a published 40-question benchmark of free perfume MCP servers run on 20 October 2026, with a median response time of 232 ms. The tool is MIT-licensed and runs as a Cloudflare Worker, which means edge caching and no single-server bottleneck. A stdio package is also available via npm and PyPI for clients that start local processes rather than connecting to remote HTTP endpoints. The REST API with OpenAPI 3.1 documentation is available separately for teams that want to call the index outside of an MCP context. ## Who should use it A developer building a fragrance recommendation chatbot, a perfume retailer adding an AI assistant to their site, or an enthusiast who wants Claude to give genuinely informed answers about note pyramids and dupe alternatives. The zero-friction setup (one URL, no account) makes it practical even for non-technical users who just want smarter perfume answers inside Claude or ChatGPT. Anyone building a multi-category shopping agent or a general product-search tool should look elsewhere. The index covers fragrance only, and the tools are read-only with no cart, checkout, or inventory integration. For broader AI workflow automation with agent coordination across domains, Orkas is a better fit, since it lets a Commander agent delegate to specialist agents across arbitrary task types rather than being scoped to a single product vertical. ## FAQ Q: Does it cost anything to use? A: No. The server is free, MIT-licensed, and requires no API key or account. The listed price is 0 EUR with no paid tiers published. Q: How many perfumes does it cover? A: The index contains 6,000+ perfumes across 249 houses, with 481 full editorial profiles (including longevity, projection, and perfumer credits) and the remainder as core profiles with notes, accords, and character descriptions. Q: Which AI clients can connect to it? A: Claude (web and Desktop), ChatGPT, Cursor, VS Code, Gemini CLI, Windsurf, Cline, Goose, and Zed all work with the single HTTP endpoint. A stdio package on npm and PyPI covers clients that start local processes. Q: How accurate is it? A: In a published 40-question benchmark of free perfume MCP servers run on 20 October 2026, it answered 39 of 40 correctly with a median response time of 232 ms, making it the top scorer in that test. Q: Can it place orders or check live stock? A: No. All 15 tools are read-only. The server returns fragrance knowledge and links to 10ml.com pages but has no write access, no cart integration, and no real-time inventory data. Q: Is there a REST API for non-MCP use? A: Yes. A REST API with OpenAPI 3.1 documentation is available separately at 10ml.com/perfume-api for teams that want to query the index outside of an MCP client.