Leadsets for GTM engineering agencies

Where data gets in the way of agency builds

A GTM engineering, RevOps or growth agency spends most of its time on systems, not lists: enrichment pipelines in n8n or Make, Clay tables, signal workflows, CRM hygiene in HubSpot or Salesforce, and handoffs to a sequencer. The data layer underneath tends to cause the same problems on every client:

• Credits make retainers hard to price. When every lookup draws down a balance, a busy month for one client changes your margin.
• Stale or unverified emails. A pipeline that writes unverified addresses into a CRM creates bounce problems you end up owning.
• Per-seat pricing. Systems run on service accounts and schedules, not on people logged into a dashboard.
• Juggling vendors. Search from one tool, emails from another, verification from a third and scraping from a fourth means four keys, four bills and four failure modes per client.

How Leadsets fits each job

Leadsets is a B2B data API and remote MCP server with eight tools: search_people, search_companies, enrich_person, enrich_company, verify_email, scrape_company, scrape_jobs and get_usage. Every tool works over both MCP and REST, so the same data layer serves a scheduled n8n flow and a client's Claude or ChatGPT agent.

Pipelines. Call https://api.leadsets.ai/v1 from an n8n HTTP node, a Make HTTP module, a Clay HTTP API column or your own code with the TypeScript and Python SDKs. Paid operations accept an Idempotency-Key, so a retry returns the first result and never counts records twice. Every response carries a usage object and a request_id, which helps when you are debugging a client's flow months after handover. Breaking changes get 90 days' notice by email.

Verified emails. Every email from enrich_person has already been through a verification waterfall (syntax, DNS and MX, a mailbox-level check, a second opinion on uncertain results) and is labelled verified, catch-all, unknown or invalid with a confidence score. Catch-all and unknown are never rounded up to valid, and invalid emails are never served as contacts. Your workflow can branch on the status field instead of guessing.

Freshness and receipts. Each field carries its source, last-checked date and confidence, and records are re-checked at least monthly. Search results include match reasons, so a client can see why a person landed in the list.

Signals and research. scrape_jobs turns a careers page into open roles with titles and dates, and scrape_company turns a homepage into products, positioning and customers named. Scraping is a paid-plan add-on billed in units (one unit is one standard page): $9 for 1,000, $19 for 5,000 or $49 for 20,000.

Skills. Leadsets publishes 50 free, MIT-licensed SKILL.md playbooks. Several map directly to agency work: n8n inbound lead enrichment, Clay table enrichment, HubSpot email backfill, Salesforce title refresh, stale record audit, companies hiring SDRs, signal alerts to Slack, usage planner, batch job runner and a multi-client ICP runner. Browse them in the full skills list.

Example: a hiring-signal pipeline for a client

Stackfold, a fictional GTM engineering agency, is building a workflow for Maya Chen, VP Revenue Operations at Northwind Analytics (northwind.co). Maya wants new accounts that are hiring sales roles to land in HubSpot each week with a verified contact and a Slack alert. Stackfold builds it in n8n:

1. A weekly trigger calls get_usage to confirm there is room in the current windows. It costs nothing.
2. search_companies pulls US B2B SaaS accounts in the 50 to 200 employee band.
3. scrape_jobs reads each careers page and keeps accounts with an SDR or RevOps role posted in the last 30 days.
4. search_people finds revenue leaders at those accounts, with match reasons.
5. enrich_person returns a current role and a verified work email when one exists. Verified contacts go to HubSpot; catch-all contacts go to a separate list for lower-volume sending; unknown and missing are logged, not sent.
6. The signal alerts to Slack skill posts each new hiring signal to Maya's channel.

Illustrative cost math: if the flow reads one careers page for each of 400 accounts a week, that is about 1,600 scrape units a month, covered by a $19 pack of 5,000 units. Search and enrichment run against the plan's flat limits, so a heavier week does not create a new bill. It hits a 429 with a Retry-After time instead, and the blocked call counts against nothing.

Which plan usually fits

All prices are flat monthly fees in USD, with no credits. Limits are published per rolling five hours and per week, and records only count when data comes back. See limits and costs for how the counters work.

• Free (1,000 enrichment records and 100 search records, 10 per call, once, with no card) is enough to prototype a flow and test filters. Scraping is not available on Free.
• Solo suits a pilot or a single small client workflow.
• Pro is the usual fit for an agency running several steady enrichment and signal workflows: 75,000 enrichments and 37,500 search records a week leaves room for backfills.
• Team fits when weekly enrichment regularly exceeds 75,000, for example large CRM backfills, and it adds priority email support within 4 hours.
• Max is for always-on agents at peak throughput.

Compare all plans on the pricing page.

Client accounts, keys and fair use

Leadsets allows one account per company, with team members added to the same account. The fair use policy does not allow sharing one key across separate companies or customers, and embedding Leadsets data in a product you sell needs a separate written agreement. Leadsets does not describe client sub-keys today. In practice, a system you hand over to a client should run on the client's own Leadsets account and key, which also keeps billing and usage clean for them. If you run work for several clients on retainer, read the terms and confirm your setup with legal@leadsets.ai before you scale it.

What Leadsets does not do yet

• No phone numbers. If a client's motion depends on dialing, use a vendor that sells phone data alongside Leadsets.
• SOC 2 is planned, not done. SOC 2 Type I readiness is underway with a target of Q3 2027. Clients whose procurement requires a SOC 2 report today will need a vendor that already has one. Current measures are on the compliance page.
• Not a CRM, a sending tool or a visual workflow builder. Leadsets feeds HubSpot, Salesforce, Instantly, Smartlead and similar tools, and is called from n8n, Make or Clay rather than replacing them.
• Business data only. No consumer data, no sensitive categories and no home addresses. US personal emails are available on paid plans.

Go deeper

• Leadsets for GTM engineers
• Leadsets for outbound agencies
• Leadsets in Clay
• GTM skills library
• Limits and costs

More who it’s for

• A B2B data API and MCP server for the agents you build
• Leadsets for Clay agencies: flat-priced data for your HTTP columns
• B2B leads for startups, from the agent you already use
• Leadsets for GTM engineers and operators
• Leadsets for outbound agencies