Model Context Protocol

Klarix Intelligence Engine MCP Server

Institutional-grade intelligence for Cursor, Claude Desktop, Windsurf, and AI agents.

Your agent stops guessing about a market and starts citing it. Fifteen tools, live public web evidence, a source URL behind every line.

Connect in one file See the 15 tools

Find your vertical in the index · Same tools over REST

CursorClaude DesktopWindsurfCustom agents

Cited or it does not ship

Every factual line carries an [S#] marker resolving to a public source URL. A buyer can audit any claim.

Markdown and typed JSON

Each tool returns rendered markdown plus a validated object, so agents chain results without re-parsing prose.

Analysis, not a contact database

No tool accepts a person name or emits an email or phone number. Roles and functions only.

Fifteen tools your agent can call

Pick a tool to see its input schema, what comes back, and a real output shape. Ten of the fifteen accept your own company, domain, and differentiators, so talk tracks come back written for your seller.

find_matched_prospectsLive

Discovers high-fit B2B prospect companies in sub-100ms using Voyage 4 semantic vector search across indexed market universes.

Input schema

Field	Type	Req
`icp_or_query` ICP description, criteria, tech stack, or problem statement	string	yes
`vertical` Filter by "AI SaaS", "Recruiting", "Manufacturing", etc.	string	no
`limit` Max prospects to return (1-25, default 10)	number	no
`min_similarity` Cosine threshold (default 0.35)	number	no
`exclude_domains` Domains to exclude from results	string[]	no
`sub_category` Narrower slice, e.g. "Serverless GPU Cloud", "Executive Search"	string	no
`headcount_min` Drops accounts whose headcount band tops out below this	number	no
`headcount_max` Drops accounts whose headcount band starts above this	number	no
`revenue_min` Annual revenue floor in USD. Advisory — the index carries no revenue field	number	no
`revenue_max` Annual revenue ceiling in USD. Advisory, same as revenue_min	number	no
`countries` Up to 10. Matched against HQ location	string[]	no
`required_tech` Up to 10. Keeps accounts with at least one in the confirmed stack	string[]	no
`our_company_name` Curates the output toward how you win	string	no
`our_domain` Your domain, e.g. "klarix.ai"	string	no
`our_differentiators` What you claim to win on	string[]	no
`client_id` Loads your saved company context, ICP, and win themes	string	no

What comes back

Ranked lookalike company list with verified domains and verticals
What they do, business model, target buyer, and confirmed tech stack
Recent growth triggers and market signals
Immediate next moves for fit scoring and battlecard generation

Ask your agent

“Find me 10 AI infrastructure companies with Python/Rust stacks serving enterprise ML teams.”

Sample OutputMarkdown & Tables

Prospect Universe — Matched Accounts (5 found)

Query: "AI infrastructure companies with Python/Rust stacks" | Universe: 150,443 accounts indexed | Latency: 75ms

Company & Domain	Vertical	Fit %	What they do	Key Tech
1. Modal (`modal.com`)	AI SaaS & DevTools	92%	Serverless cloud platform for running AI/ML models	Python, Rust, CUDA
2. Baseten (`baseten.co`)	AI SaaS & DevTools	88%	High-performance model inference infrastructure	Truss, Python, TensorRT
3. Together AI (`together.ai`)	AI SaaS & DevTools	85%	Cloud platform for building and running open models	PyTorch, vLLM, Triton

Recommended next moves:

score_prospect_fit(company="Modal", domain="modal.com")
get_competitor_battlecard(competitor_domain="modal.com")

Six verticals in the index

Every company in the vector universe carries one of these labels. Find yours to see the sub-segments it covers, a query that returns real matches, and the tool chain we benchmarked against that buyer.

Pass the value verbatim as vertical to scope a search to one slice. Omit it to search all 150,443 indexed companies.

AI SaaS & DevTools

LLM infrastructure, MLOps, agent frameworks, developer platforms, and cloud tooling.

LLM & vector infra
MLOps and data tooling
Agent frameworks
API & cloud platforms

vertical

AI SaaS & DevTools

“vector database companies selling to enterprise RAG teams”

Benchmarked chain · VP RevOps

Find lookalikes, read their stack, then write the displacement plan.

find_matched_prospects
teardown_tech_stack
get_buying_committee
generate_displacement_playbook

Precision Manufacturing & Hardware

Contract manufacturers, EMS, and hardware OEMs where the spec sheet is the sales argument.

PCBA and box build
CNC & injection molding
Battery pack assembly
Aerospace and robotics

vertical

Precision Manufacturing & Hardware

“contract manufacturers building custom battery packs for medical OEMs”

Benchmarked chain · VP Commercial Ops

Win a spec-driven bid against a named incumbent.

find_matched_prospects
analyze_head_to_head
search_scientific_evidence
teardown_product_spec

Recruiting & Staffing

Retained search, technical staffing, and talent advisory firms selling into hiring teams.

Retained executive search
Technical & engineering staffing
PE and board search
RPO

vertical

Recruiting & Staffing

“boutique executive search firms placing VP Engineering roles at Series B startups”

Benchmarked chain · Managing Partner

Catch a hiring trigger and reach the committee before the req posts.

detect_account_triggers
get_buying_committee
generate_outreach_sequence

Fintech & Financial Services

Payments, embedded finance, billing, treasury, and financial data infrastructure.

B2B payments & embedded finance
Billing and subscriptions
Treasury & spend
Risk and compliance

vertical

Fintech & Financial Services

“embedded payments platforms serving vertical SaaS companies in North America”

Benchmarked chain · Head of Partnerships

Qualify the account, then out-position the incumbent processor.

score_prospect_fit
get_competitor_battlecard
generate_displacement_playbook

Healthcare & Life Sciences

Medical devices, diagnostics, clinical software, and biotech tooling where evidence gates the deal.

Medical device OEMs
IVD & diagnostics
Clinical and trial software
Biotech tools

vertical

Healthcare & Life Sciences

“in vitro diagnostics companies commercializing FDA-cleared point-of-care assays”

Benchmarked chain · CTO

Back a clinical claim with citable literature before the committee meeting.

get_deep_research
search_scientific_evidence
get_buying_committee

B2B Services & Enterprise

Enterprise software, cybersecurity, GRC, agencies, and professional services.

Cybersecurity & IAM
GRC and compliance automation
Enterprise ops software
Agencies & consultancies

vertical

B2B Services & Enterprise

“cloud security platforms displacing legacy SIEM in mid-market accounts”

Benchmarked chain · Field CISO

Map the category, then take a live head-to-head into the room.

teardown_tech_stack
generate_swot_analysis
analyze_head_to_head

One file, then reload

Paste the block into your client config, add your key, restart. The server starts even with keys missing so you can list tools first and read klarix://status to see what resolved. No MCP client? The same fifteen tools answer over REST.

File

.cursor/mcp.json (project) or ~/.cursor/mcp.json (global)

{
  "mcpServers": {
    "klarix-intelligence": {
      "url": "https://mcp.klarix.ai/mcp",
      "headers": {
        "Authorization": "Bearer klx_live_your_token_here"
      }
    }
  }
}

Paste, reload Cursor, then check Settings → MCP for fourteen green tools. Older builds that reject a "url" server need the legacy SSE block on the CLI / HTTP tab.

Token access plans

One query means one tool call. Research and synthesis costs are already inside the price, so a deep call and a quick one cost you the same.

Free Trial

$0 / 3 queries

3 free queries

Test all 15 tools live in Cursor or Claude Desktop before committing.

All 15 live intelligence tools
3 full-depth tool queries included
Instant access via npx @klarix/mcp-server
Markdown plus typed JSON output
Zero credit card required

Most picked

Pro

$197/mo

1,000 queries/mo

For sales engineers, founders, and revenue operators querying competitive intel daily.

Everything in Free Trial
1,000 tool queries per month
Deep Research depth and scientific retrieval
Priority synthesis queue
Spoken talk tracks and 2-move battlecards

Team

$497/mo

5,000 queries/mo

For multi-rep sales teams, GTM agencies, and automated workflow orchestrations.

Everything in Pro
5,000 tool queries per month
Shared org token and multi-seat access
Hosted SSE endpoint (mcp.klarix.ai)
Webhooks for long-running batch jobs

Included with every Klarix CI Retainer

Every Klarix Done-For-You CI client ($2,997–$8,997/mo) receives full MCP server access for their internal engineering, sales, and strategy teams. Run real-time prospect fit scores, competitor battlecards, and SWOT checks directly in Cursor and Claude Desktop alongside your monthly branded PDF bundles.

Unlimited internal team access across Cursor, Claude Desktop, and Windsurf
Shared workspace token with custom rate limits and usage visibility
Powered by Klarix intelligence infra — zero provider API keys to manage
Direct synchronization with your monthly researched account portfolio

Past what an agent can carry

Need 25 to 100+ fully managed PDF bundles, scored decision-makers, and monthly compounding refresh?

The MCP server answers one question at a time. Klarix CI-as-a-Service runs the whole market for you: ranked dossiers, battlecards, SWOT, contacts, and outreach delivered as branded PDFs in 3 to 7 days, refreshed every month.

See CI-as-a-Service pricingFrom $2,997/moSee what a bundle looks like