The Best PulseMCP Alternative for Developers in 2026

PulseMCP is an MCP server directory that tracks 21,690+ servers updated daily, filterable by classification (official providers, community, Anthropic references), remote availability, popularity, and release date. Each listing shows an estimated visitor count, a popularity ranking on a global leaderboard, release date, transport type (such as Streamable HTTP), auth requirements, and where available a standardized server.json file with installation instructions and a copy-paste remote endpoint URL. The directory covers everything from Figma, Notion, and GitHub to niche servers for camping searches and roto-mold layout feasibility. However, PulseMCP is currently in a rework phase: new server submissions and listing changes are paused. It also publishes ecosystem content, including guest posts from MCP founders on pricing and product strategy.

pricing See pricing

• + Massive breadth: 21,690+ servers tracked and updated daily
• + Useful metadata per listing: transport type, auth type, cost flags, estimated visitors, release dates
• + Standardized server.json files make installation straightforward for developers
• + Free to browse, and many listed servers are flagged as entirely free to use
• + Publishes ecosystem editorial content, including founder guest posts on MCP pricing

• − New server submissions and listing changes are currently paused during a rework of the ingestion system
• − No benchmarking of listed servers — nothing tells you whether a server actually improves agent performance
• − No security scanning or vulnerability checks on listed servers
• − Estimated visitor numbers are algorithmic approximations, not verified usage data
• − Sheer volume (thousands of low-traffic listings) makes quality hard to judge

https://www.pulsemcp.com

SkillzBench is AIQ Labs' free, independent directory of AI agent skills plus a free MCP tool that connects any agent to the skills you choose. Agent skills (SKILL.md instruction packages for Claude Code, Codex, Cursor, Gemini CLI, and chat-app agents) are everywhere, but published research found roughly 26% of skills carry at least one vulnerability — and SkillzBench is being built to solve exactly that. Its pipeline: Discover (continuously crawl and dedupe skills from GitHub), Verify (pin every version to a commit SHA, validate, license-check, security-scan, quality-judge), Benchmark (test every skill in a sandbox with the same model, with and without the skill, and publish the Lift in percentage points), Equip (one free MCP server plus per-agent loadouts where the agent only sees what was chosen), and Stay current (track upstream changes, re-scan, re-benchmark, notify). A 0-100 trust score combines security, quality, license, maintenance, and community signals; Lift is always shown separately from ratings and money never touches scores. SkillzBench is pre-launch — the waitlist is live, the product is not yet public — and everything on the platform will be free, funded by advertising around blog articles.

pricing 0

• + Everything free: the directory, the MCP connector, benchmarks, and alerts
• + Directly addresses the documented risk that ~26% of skills carry at least one vulnerability
• + Published Lift benchmarks show whether a skill actually helps, separately from ratings
• + Per-agent loadouts mean each agent only sees the skills and tools chosen for it
• + Commit-SHA version pinning plus ongoing re-scanning keeps skills trustworthy over time
• + Money never touches scores — advertising revenue is kept separate from listings

• − Pre-launch: the product is not yet public, so none of it can be used today
• − Focused on agent skills rather than the full universe of MCP servers PulseMCP indexes
• − No published benchmark results or user counts yet, by design, until launch

https://skillzbench.com

MCP Hive takes the opposite approach to a directory: instead of browsing thousands of listings, you get access to all participating MCP servers through a single subscription. The Starter plan at $19.99/month includes 8,000 MCP calls (roughly $0.0025 per call), Professional at $39.99/month includes 16,000 calls plus a $20/month credit for priced MCP servers, and Enterprise starts at $249.99/month with pay-as-you-go beyond the included calls at $20 per additional 8,000 calls, a custom spending cap, and enterprise-level customer support. All plans include a 30-day free trial with no credit card required, and you can cancel anytime. On the provider side, MCP Hive retains 30% of what a server earns and pays out 70%. Per-call costs for priced MCP servers are shown during discovery before a call is made, and the console tracks calls used and remaining each billing period.

pricing 19.99/month · free trial

• + One subscription and one account replaces managing many individual MCP server API keys
• + Clear, published pricing with no setup fee and no separate per-call fee for included servers
• + 30-day free trial on all plans with no credit card required
• + Enterprise tier adds pay-as-you-go scaling under a custom spending cap
• + Provider-side economics (70/30 payout split) are disclosed openly

• − Costs scale with usage — heavy agent workloads can push you up plans quickly
• − Access is limited to participating MCP servers, not the full 21,690+ breadth of a directory like PulseMCP
• − No published benchmarking or security verification of the servers it carries

https://mcp-hive.com

Ref is possibly the first standalone MCP server built as a paid product from the ground up, providing precise documentation search for AI coding agents. Its two tools — ref_search_documentation and ref_read_url — each consume one credit per call, letting coding agents reference accurate documentation and make fewer mistakes without wasting context window. Users can search a public index of library and package documentation maintained by a custom web crawler, and build private indexes of PDFs and internal GitHub repos. Its credit-based pricing gives 200 free credits that never expire, then $9/month for 1,000 credits ($0.009 per search). Since launching this pricing roughly three months before publication, Ref reported thousands of weekly users and hundreds of subscribers. Its cost structure spans search infrastructure (Turbopuffer), LLM retrieval-pipeline costs, and web crawling — its highest expense. It positions itself between free Context7 and paid search APIs like Tavily and Exa.

pricing 9/month · free trial

• + 200 free credits that never expire let you evaluate before paying
• + Simple, low entry price at $9/month for 1,000 searches
• + Private indexes of team documentation alongside the curated public index
• + Proven traction: thousands of weekly users and hundreds of subscribers reported within three months of launch
• + Designed specifically to reduce coding-agent mistakes from stale or hallucinated docs

• − Single-purpose — only documentation search, not a broad MCP ecosystem
• − Costs scale at $0.009 per search for high-volume agents
• − Web crawling is its highest expense, creating pricing pressure as the index grows

https://ref.tools

Context7 is a free MCP server providing basic RAG (retrieval-augmented generation) for documentation, drawing primarily from public GitHub repositories. It is widely cited as the free baseline in the coding-agent documentation-search space: Ref's founder positioned Ref explicitly between Context7's free basic RAG and paid search APIs like Tavily and Exa, which charge roughly $0.01 per search. For developers who want documentation lookup inside their coding agent without any subscription or credit system, Context7 is the zero-cost entry point. The tradeoff is precision and scope — it's described as basic RAG rather than premium search, and it focuses on public repos rather than private team documentation indexes. For small businesses testing whether documentation MCP servers help their agents at all, it's the cheapest way to find out.

pricing 0

• + Completely free — no credits, no subscription, no card
• + Zero-friction way to test documentation lookup in a coding agent
• + Serves as the ecosystem's free baseline for docs RAG

• − Basic RAG quality — positioned below premium search options on precision
• − Limited to primarily public GitHub repo documentation
• − No private team documentation indexing

Tavily is an agent-focused search and scraping API with a popular MCP server, charging roughly $0.01 per search with volume discounts. Along with Exa, it's one of the two paid search APIs most commonly used as a pricing benchmark in the agent ecosystem — Ref's founder cited both when positioning a paid documentation-search MCP server between free Context7 and paid general search. For developers building agents that need live web search and page scraping beyond static documentation indexes, Tavily's MCP server is a mainstream, widely integrated option. It appears in PulseMCP's trending server list alongside other high-traffic servers like Figma, Notion, and GitHub. The per-search model means costs scale directly with agent query volume, which suits variable workloads but requires monitoring for high-volume agents.

pricing 1 tiers

• + Broad live web search and scraping, not limited to documentation indexes
• + Well-established in the agent ecosystem with a popular MCP server
• + Volume discounts for heavier usage

• − Roughly $0.01 per search adds up for agents making thousands of queries
• − General-purpose search is less precise for documentation lookup than dedicated docs servers
• − No published skill-level benchmarking or security verification