The 5 Best PulseMCP Alternatives in 2026

PulseMCP is a large MCP server directory aimed at people choosing servers for their MCP clients. Its catalog runs to thousands of entries (individual pages show counts around 3,855–3,856 filtered listings, with the site branding itself as 21,690+ updated daily), each with classification (official vs community), estimated visitor or download counts, popularity rankings, release dates, remote endpoints, auth requirements and transport details such as Streamable HTTP. Some listings include a standardized server.json file with installation instructions and configuration options, and several servers are marked as entirely free to use. PulseMCP also publishes ecosystem content, including guest posts from MCP builders such as Ref's founder on pricing a paid MCP server. Its current limitation is operational: the site displays a notice that new server submissions and listing changes are paused while it reworks how it ingests and manages listings. For small businesses, PulseMCP works as a research and discovery layer — you browse, compare and then configure servers yourself in your MCP client.

pricing See pricing

• + Very large catalog of MCP servers with official-provider and remote-available filters
• + Per-server detail pages include exact remote endpoint URLs, auth type and transport protocol
• + Popularity rankings and visitor estimates help gauge which servers are actually used
• + Free to browse for discovery and research

• − New server submissions and listing changes are paused while the ingestion pipeline is reworked
• − No security scanning, license checks or benchmarked performance scores for servers
• − It is a directory, not an access layer — you still configure each server in your client yourself
• − Visitor figures are algorithmic estimates, not verified usage numbers

https://www.pulsemcp.com

MCP Hive takes a different approach from a directory: it is a subscription access layer for MCP servers. One plan unlocks all participating MCP servers, with monthly call allowances and credits for servers that charge per call. The Starter plan is $19.99/month for 8,000 MCP calls (roughly $0.0025 per call), Professional is $39.99/month for 16,000 calls plus a $20/month credit for priced MCPs, and Enterprise starts at $249.99/month with pay-as-you-go beyond credits — $20 per additional 8,000 calls — under a custom spending cap. All plans include a 30-day free trial with no credit card required, and you can cancel anytime. Priced MCP servers show their per-call cost during discovery before a call is made, so costs are visible upfront. On the supply side, MCP Hive retains 30% of what a server earns and pays out 70%. A console tracks calls used and remaining per billing period, and Enterprise-level customer support is included on the top plan.

pricing 19.99/month · free trial

• + One subscription instead of configuring and paying for servers individually
• + Transparent per-call economics (~$0.0025/call on Starter) and upfront per-call pricing for paid MCPs
• + 30-day free trial on every plan, cancellable anytime
• + Enterprise plan supports heavy agent workloads with pay-as-you-go and a spending cap

• − Call allowances can bind: Starter's 8,000 calls/month is modest for autonomous agents
• − Pay-as-you-go beyond credits is only available on the $249.99+/month Enterprise plan
• − Limited to participating MCP servers — coverage depends on the catalog
• − Platform takes a 30% cut of server earnings, which may affect which servers participate

https://mcp-hive.com

Ref is a premium documentation-search MCP server built specifically for AI coding agents, and one of the first standalone paid MCP servers. It provides two tools — ref_search_documentation and ref_read_url — that let coding agents search a curated public index of library and package documentation and scrape pages on the fly, so agents reference accurate docs instead of hallucinating. Its differentiators are search precision optimized to avoid wasting context window, and the ability to build private indexes of your team's documentation, PDFs and private GitHub repos alongside the public index. Pricing is credit-based: 200 free credits that never expire, then $9/month for 1,000 credits (about $0.009 per search); you pay only for searches, not indexing. Since launching this packaging three months before its founder's guest post, Ref reported thousands of weekly users and hundreds of paying subscribers. Small teams can share a single API key across the whole team, which keeps costs down for collaborative use.

pricing 9/month · free trial

• + Search precision specifically optimized for coding agents and context-window efficiency
• + Private team indexing is a real differentiator over free documentation tools
• + Low entry cost at $9/month with 200 free non-expiring credits to start
• + Trivial team rollout — share one API key across the team

• − Narrow scope: documentation search only, not a general MCP access or skills platform
• − Credit costs scale with agent volume — thousands of daily agent searches add up
• − Positioned between free Context7 and Tavily/Exa, so value depends on how much precision matters to you
• − Pricing figures come from a founder guest post, not a public pricing page in this research

https://ref.tools

Context7 provides a free MCP server with basic RAG (retrieval-augmented generation) for documentation, primarily sourced from public GitHub repositories. In the competitive landscape described by Ref's founder, Context7 occupies the free end of the documentation-search market: it gives coding agents access to library and package docs without any per-search cost, making it the default starting point for developers who want their agents to stop hallucinating outdated APIs but don't yet need premium precision or private indexing. Because it draws mainly from public repos, it suits open-source-heavy workflows — individual developers, hobby projects, and small teams working with well-documented public libraries. Its limitations are the flip side of its price: search precision is positioned below Ref's documentation-specialized retrieval and below the paid agent-search APIs, and there is no capability for indexing your team's private documentation. For small businesses evaluating MCP documentation tools, Context7 is the zero-cost baseline against which paid options like Ref (~$0.009/search) and Tavily/Exa (~$0.01/search) position themselves.

pricing 0

• + Free — no per-search cost, unlike Ref, Tavily and Exa
• + Easy baseline for coding agents that need public library docs
• + Widely referenced as the free standard in the MCP documentation-search category

• − Basic RAG only — lower search precision than Ref's specialized retrieval
• − Primarily public GitHub repos; no private-index capability for team documentation
• − No volume guarantees or premium support implied by the free positioning

Tavily offers agent-focused search and scraping services with a popular MCP server, charging roughly $0.01 per search with volume discounts. In the MCP ecosystem it serves as one of the two standard paid search benchmarks (alongside Exa) that newer entrants like Ref price themselves against. Where Ref specializes in precise documentation retrieval for coding agents, Tavily is a general-purpose search layer for agents: it handles web search and page scraping so agents can pull in current information from across the web rather than from a curated documentation index. That breadth is its strength — small businesses building agents that need live web data (research, monitoring, content gathering) get a managed, agent-optimized search API without running their own crawler. The trade-offs relative to documentation specialists are precision on technical docs and cost predictability: at roughly $0.01 per search, heavy autonomous agents running thousands of queries accumulate meaningful monthly costs, though volume discounts soften this at scale.

pricing See pricing

• + General web search and scraping in one agent-ready API
• + Popular, established MCP server in the ecosystem
• + Volume discounts available for heavier usage

• − Roughly $0.01 per search adds up quickly for autonomous agents making thousands of calls
• − General search rather than documentation-specialized precision retrieval
• − Per-call pricing makes monthly costs less predictable than flat subscriptions

Exa is the other major paid agent-search benchmark in the MCP ecosystem, offering agent-focused search and scraping services at roughly $0.01 per search with volume discounts, plus its own MCP server. Like Tavily, it gives AI agents a managed way to search the web and pull in pages, and it is cited alongside Tavily as the pricing reference point that documentation-search specialist Ref positioned itself against. For small businesses, Exa fits workflows where agents need broad, current web information — competitive research, market monitoring, content sourcing — rather than precise lookups in a curated documentation index. The considerations mirror Tavily's: per-search pricing is flexible but can scale steeply with autonomous agent volume, and general web search trades precision on technical documentation for breadth. Choosing between Exa and Tavily in this research comes down to evaluation on your own queries, since both occupy the same price band (~$0.01/search) and the same functional category of agent-oriented search with MCP connectivity.

pricing See pricing

• + Established agent-search API with MCP connectivity
• + Same accessible price band as Tavily (~$0.01/search) with volume discounts
• + Broad web coverage for research-style agent tasks

• − Per-search costs compound quickly for high-volume autonomous agents
• − Not specialized for documentation precision the way Ref is
• − Research here provides no differentiating detail versus Tavily — test both on your own queries

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 — have proliferated, but published research found roughly 26% of skills carry at least one vulnerability, and nobody could tell which ones actually help. SkillzBench is being built around five pillars: Discover (continuously crawling and deduplicating skills from GitHub), Verify (pinning every version to a commit SHA, validating, license-checking, security-scanning and quality-judging), Benchmark (testing every skill in a sandbox with the same model, with and without the skill, and publishing the Lift in percentage points), Equip (one free MCP server plus per-agent loadouts, where you toggle skills and tools per agent and the agent only sees what you chose), and Stay current (tracking upstream changes, re-scanning, re-benchmarking and notifying you). Every skill gets a 0-100 trust score from security, quality, license, maintenance and community signals; Lift is always shown separately, and money never touches scores. Everything is free — the directory, MCP connector, benchmarks and alerts — funded by blog advertising that never influences scores or listings. The product is pre-launch: the waitlist and newsletter are live.

pricing 0

• + Completely free — directory, MCP connector, benchmarks and alerts, funded by blog advertising that never touches scores
• + Addresses a real safety gap: ~26% of published skills carry at least one vulnerability
• + Benchmarked Lift scores show whether a skill actually helps, not just whether it exists
• + Per-agent loadouts give precise control over what each agent can access
• + Continuous upstream monitoring keeps trust scores current as skills change

• − Pre-launch — the product is not yet public; the waitlist/newsletter is live
• − Focused on agent skills rather than being a general MCP server marketplace
• − No live user counts or published benchmark results yet, given the pre-launch stage

https://skillzbench.com