SalesHub · agentic GTM data

The data department your AI sales agent was missing.

411M verified contacts. 78M companies. Technographics across 120K+ tools. Real-time intent signals. Exposed as an MCP server so your AI agent queries B2B data the way it queries any other tool — at flat-rate unlimited pricing, with a sub-2% email bounce rate.

saleshub — codex
enrichbuild a listverify
SALESHUB
saleshub mcp·v2.4·connected to salestools.io graph live
411,204,883 contacts·78,142,990 companies·20+ source waterfall
>find Heads of RevOps at US Series-B SaaS, 50–500 staff, running Salesforce
saleshub.searchtitle~RevOpsstage:series_btech:salesforce
scanning firmographics + technographics88%
companies614
contacts1,872
with mobile1,403
net-new1,560
1,872 verified contacts·streamed to your CRM
411M
verified contacts — emails, direct dials, LinkedIn
78M
companies — firmographics, funding, headcount
<2%
email bounce rate — real-time verification
Flat-rate
unlimited queries — no credits, no per-record fees

How it works

Three steps from first prompt to production.

01

Connect

Run `claude mcp add saleshub` in your terminal. Your agent — Claude Code, Codex, Salestools GTM, or any MCP-aware client — gets structured access to every SalesHub dataset in under a minute.

02

Ask in natural language

Your agent issues real prompts: 'enrich these accounts and find the buyers', 'who at Ramp is hiring senior platform engineers this quarter', 'show me B2B SaaS companies in EU with Snowflake in their stack'.

03

Ship outreach

Verified contacts, citations, and confidence scores flow back as structured JSON. Your agent drafts outreach, your CRM stays clean, your reply rates go up — at flat-rate cost.

Built for every seat

However you go to market, the agent adapts.

Founders building AI sales agents

Skip building a contact pipeline. SalesHub gives your agent verified emails, mobiles, and intent signals out of the box — the data layer you'd otherwise spend a year and seven figures to assemble.

GTM teams running autonomous outbound

Replace four overlapping tools (Apollo, ZoomInfo, Crunchbase, an intent vendor) with one MCP server. Flat-rate pricing means you actually let the agent run at full volume.

RevOps building data pipelines

Stop reconciling six aggregator schemas. SalesHub is a single source of truth for accounts, contacts, technographics, and signals — versioned, structured, agent-callable.

AI product teams shipping agents

Embed B2B data inside your own product. SalesHub's MCP server gives your agent the same lookup power as a senior SDR — without the cost of building or licensing a data warehouse.

Plugs into your stack

Native integrations across the tools your team already uses.

Claude CodeCodexCursorSalestools GTMGrain agent fleetOpenAI MCP clientAnthropic MCP clientSalesforceHubSpotOutreachSalesloftCustom MCP clients

Frequently asked

How is SalesHub different from Apollo, ZoomInfo, or Clay?
Those products are built for humans clicking through dashboards, with per-record pricing that punishes autonomous use. SalesHub is built for AI agents — it ships as a native MCP server your agent calls directly, with flat-rate unlimited pricing so volume is never the constraint. The dataset is bigger (411M contacts, 78M companies) and the bounce rate is lower (<2%) because we re-verify continuously.
What does the dataset actually cover?
Verified business and personal email, mobile and direct-dial phone, LinkedIn URL, role, seniority, full company firmographics (revenue, headcount, funding, HQ, locations), technographics across 120K+ technologies, and real-time intent signals (hiring spikes, job changes, funding rounds, content publication, technology adoption). All structured, all citable, all callable through one MCP server.
How does the MCP integration work?
One command: `claude mcp add saleshub`. That registers SalesHub as an MCP server with your agent runtime (Claude Code, Codex, Cursor, custom clients). Your agent then queries SalesHub using natural language or structured tool calls — `search_prospects`, `enrich_company`, `get_phone`, `get_email`, and friends — and receives JSON back with verified data and provenance.
What does it cost?
Flat-rate, unlimited. No credit counting, no per-record fees, no overage. One price covers all data volume — built specifically so AI agents can run at autonomous scale without the cost spiraling. Pricing tiers depend on team size and use case; talk to sales for a quote.
How fresh is the data?
Companies and intent signals refresh daily from primary sources. Verified contacts are re-verified continuously to keep email bounce rate below 2% — the lowest in the industry. Stale records are flagged with provenance so your agent knows the difference between yesterday's hire and last quarter's exit.
Is my prospect search history used to train models?
No. Your queries, enrichment requests, and pulled contacts are never used to train models. Enterprise contracts ship with zero-retention serving by default and a Data Protection Addendum.

Give your agent the data department it deserves.

411M contacts. 78M companies. One MCP command. Flat-rate, unlimited, agent-priced.