Use case

Lead data for Cursor and coding agents

To give Cursor, Claude Code or Codex lead data, paste one line into the agent: set up https://leadmcp.ai/SKILL.md. The agent connects LeadMCP by itself and asks you to approve it once. Then it can count, search, enrich and export contacts from your code and your files.

Start free

Give this to your agent: set up https://leadmcp.ai/SKILL.md

Copy the line and paste it into your agent. All use cases

01 Connect

Connect in 2 minutes

Claude Code and Codex set themselves up from one line. Cursor can do the same, or you can add the config file. You need no API key.

Claude Code

01
Paste one line into Claude Code
Copy the line below. Your agent reads the setup, registers itself and shows you a link. No API key, no config to edit.
Give this to your agent: set up https://leadmcp.ai/SKILL.md
02
Click Approve
Open the link your agent shows and click Approve. Your agent finishes the setup and calls the guide tool by itself.

Codex

01
Paste one line into Codex
Copy the line below. Your agent reads the setup, registers itself and shows you a link. No API key, no config to edit.
Give this to your agent: set up https://leadmcp.ai/SKILL.md
02
Click Approve
Open the link your agent shows and click Approve. Your agent finishes the setup and calls the guide tool by itself.

Cursor

01
Add LeadMCP to Cursor
Open Cursor → Settings → MCP → Add new global MCP server, and paste the config into .cursor/mcp.json. No API key: Cursor opens a LeadMCP tab, you sign in once and click Approve.
.cursor/mcp.json
```
{
  "mcpServers": {
    "leadmcp": {
      "url": "https://mcp.leadmcp.ai"
    }
  }
}
```
02
Enable LeadMCP and ask Cursor to find leads
Save the file and turn the leadmcp server on. Then ask Cursor to count contacts for a segment you care about.

02 Ask

Try these prompts

Type them as they are. Each one shows the tool the agent calls and what it costs. Counts are always free.

Pull every CTO at US companies with 201 to 1K people into data/ctos.jsonl.
start_export contact credits
The agent counts first, for free. Then it starts an export with countries, seniority, functions, company_size and format set to jsonl, polls get_export and saves the file. Each exported contact costs one contact credit.
Add email and phone columns to leads.csv. Each row has a LinkedIn URL.
enrich_contact contact credits
The agent reads the file and sends the URLs to enrich_contact in batches of 100. It writes the matches back as new columns. Each stored match costs one contact credit, and a miss costs nothing.
Skip every company in customers.csv and count what is left.
count_matches free
The agent reads the domains from the file into exclude_domains and counts the rest, for free.
Find software companies founded after 2018 with 11 to 200 people.
find_companies contact credits
The agent reads the industry value with get_values, then searches companies with industries, founded_year_min and company_size. Each returned company costs one contact credit.
Write a script that runs this search every Monday through the REST API.
create_api_key free
The agent makes the account's API key in its signed-in session and writes the script. The script sends the key as a Bearer token. Making the key is free.
Watch our CEO's LinkedIn profile and log new engagers every day.
create_alert premium credits
The agent creates an alert that checks the profile daily and reads new rows with list_signals. Alerts are on the paid plan, and each new lead costs one premium credit.

03 Example

One prompt, step by step

The tool calls the agent makes for one request, and the shape of what comes back.

export US sales leaders at companies with 1 to 50 people, with an email and a phone, to a CSV

1
count_matches freecountries: ["us"]functions: ["sales_and_bizdev"]seniority: ["vp", "director", "c_level"]has_work_email: truehas_phone: truecompany_size: ["1-10", "11-50"]33,394 matches
2
start_export contact creditscountries: ["us"]functions: ["sales_and_bizdev"]seniority: ["vp", "director", "c_level"]has_work_email: truehas_phone: truecompany_size: ["1-10", "11-50"]format: "csv"job queued, with a preview of the first rows
3
get_export freecompleted, with a download link that works for 7 days

sales-leaders-us-1-50.csv 33,394 rows · the preview below

The first rows of the export
name	company	work_email	mobile_phone
Imogen ValeVP of Sales	Linear	`imogen.vale@linear.app`	`+1 415 555 0142`
Caspian RoweHead of Sales	Retool	`caspian.rowe@retool.com`	`+1 646 555 0178`
Ottilie MarshVP Sales, Americas	Sentry	`ottilie.marsh@sentry.io`	`+1 512 555 0116`

The agent spent 33,394 contact credits, one for each exported contact. The count cost nothing.

Sample rows: fictional people at real companies. The counts are real, from the live database on 6 Oct 2026.

One flat plan

Unlimited leads. One flat price.

Unlimited leads, searches and counts
$0 extra for phones
$500 a month, flat

No per-seat fees. No per-row charges. Your agent can size every list before it spends a contact credit. See pricing

Start free with 1,000 contact credits, one time. Signal tools, such as alerts and post engagers, use premium credits, which come with the paid plan.

04 Questions

Common questions

What does the one setup line do?

The agent reads the LeadMCP skill file and registers itself. It shows you a link, and you click Approve. Then it adds the MCP server and calls the guide tool. You never paste an API key.

Does it work in Claude Code and Codex too?

Yes. Paste the same line into Claude Code, Codex or Cursor. Each one connects with one Approve click.

Can a script use LeadMCP without MCP?

Yes. Use the REST API with your API key as a Bearer token. Every endpoint, filter and response shape is in the API reference.

What does it cost to find leads in Cursor?

LeadMCP is one plan at $500 a month, with unlimited leads and searches. Counts are free, and phones come at no extra cost. You can start free with 1,000 contact credits, one time. Signal tools use premium credits, which come with the paid plan. The pricing page has the details.

Comparing tools? Read LeadMCP vs Clay and Apollo alternatives for AI agents. Or see all use cases.