Describe a customer.
Get the people.

OpenLeads searches free public sources for the businesses and people you describe, finds their email addresses and tells you how it knows each one. It runs on your machine, with no API keys and nothing to pay.

Or run npx openleads. Quickstart

$ openleads find"dentists in Austin"10 leads in 41 s
Maya Okafor Owner and lead dentistNorthwind Dental[email protected]found
Daniel Reyes Practice managerNorthwind Dental[email protected]pattern
Priya Shah OrthodontistLakeline Orthodontics[email protected]found
Tom Brennan Office managerBarton Creek Family Dentistry[email protected]guessed
Front desk Contact addressSouth Congress Smiles[email protected]found
Elena Marsh Lead hygienistMueller Dental Studio[email protected]pattern
Found Pattern Guessed Deliverable Risky An example. The people and domains are invented.

Every address says how we know it.

Most finders give you one confidence score. OpenLeads tells you where each address came from, so you can decide what is worth sending.

Found

Published on the company's site, in a public commit or a security.txt file, or confirmed by the mail server.

Pattern

Built from an address format already seen at that company, like first.last or first initial plus last name.

Guessed

The most common format for that name. Nothing confirms it, and the label says so.

The color answers a second question: will it deliver? Seven checks decide, and most of them work without port 25, which home internet providers block.

Deliverable sent by default Risky held for your review Will bounce dropped
  • Mail servers Two DNS-over-HTTPS resolvers must agree
  • SPF, DMARC and provider From the domain's TXT records
  • Disposable, role and free mail Checked against lists on your machine
  • Gravatar A public profile means a real inbox
  • Published addresses Harvested from sites, commits and security.txt
  • Learned patterns Each run teaches the next one
  • SMTP and catch-all Probed twice when port 25 is open

We publish how often we guess.

Every night a GitHub Action runs the same real searches against the live sources, asks for ten leads each, and commits what came back. No cache, no cherry-picking, and the empty nights stay in.

SearchReturnedLeadsFoundPatternGuessedTime
dentists in Austin000015 s
marketing agencies in Miami1042448 s
law firms in London4103117 s
accountants in Toronto410377 s
software companies in Berlin10604103 s
gyms in Sydney440090 s
fintech founders1020818 s
AI startup founders1050517 s
rust developers in Berlin10802103 s
machine learning researchers1030751 s
emails at stripe.com42205 s
real estate agents in Chicago1090152 s

Run of 28 Oct 2026 on version 5.0.0: 11 of 12 searches returned leads, 72% of the requested leads came back, and 57% of the emails were found or built from a seen pattern. Every night's results

A sales workspace on localhost.

Run openleads web and everything you find lands in one place: people and companies with Apollo-style filters, lists, sequences, tasks, deals and analytics. It is a single SQLite file in ~/.openleads, and it never leaves your machine.

Want to look first? openleads demo opens it on an invented workspace.

The People page of the OpenLeads dashboard: a filter column, and a table of contacts with their title, company, email with its evidence mark, stage and last contact.

The People page, on the demo workspace.

Find
Plain-English search across ten free sources. Company lookups with tech stack, socials and phone numbers. Enrich any CSV.
Verify
Found, pattern or guessed on every address, seven deliverability checks, and automatic suppression of bounces and unsubscribes.
Reach out
Drafts you edit before they send. Sequences with email, call, LinkedIn and task steps. Warmup caps and one-click unsubscribe.
Track
Stages that move on their own, ICP fit from A to D, a task queue, a deal pipeline with a weighted forecast, and analytics.
Automate
Recipes and watchers on a schedule, drawn as workflows. Exports to Sheets, Notion, Airtable, HubSpot and webhooks.
Bring your data
Import Apollo, HubSpot and LinkedIn exports with the columns mapped for you, and export back in their formats.

Use it from the terminal, the browser, Python or Claude.

Terminal

Results print as a table, or as CSV and JSON you can pipe.

# ten leads, every email labeled
openleads find "accountants in Toronto"
openleads find "fintech founders" -f csv > leads.csv

Browser

The workspace on 127.0.0.1, bound to your machine only.

openleads web
# or try it on invented data
openleads demo

Python

One call returns leads you can loop over or save.

import openleads

leads = openleads.find("gyms in Sydney", count=20)
leads.write("gyms.csv")

Claude and Cursor

An MCP server that finds leads and works your lists, tasks and sequences.

claude mcp add openleads -- openleads mcp
# then ask:
# find 20 dentists in Austin and add them to a list

Compared with the free tiers.

FeatureOpenLeadsApollo and Hunter, free
Cost$0, no credit limitsCredits, then paid plans
API keyNoneRequired
Email labelsFound, pattern or guessed on every addressOne confidence score
Your own listPaste a CSV, get every email labeledPaid feature
Sequences, tasks, dealsIncluded, unlimitedLimited, then paid
Where your data livesYour machineTheir cloud
Source codeOpen, AGPL-3.0Closed

Questions people ask.

Is it really free?

Yes. It only uses free public sources and your own mailbox, so there is nothing to meter. The code is AGPL-3.0; if you want to build it into a closed-source product, there is a commercial license.

Where do the leads come from?

OpenStreetMap for local businesses, Y Combinator and Hacker News for startups, Wikidata and SEC EDGAR for companies, GitHub, Hugging Face and OpenAlex for builders and researchers, the NPI registry for US clinicians, and company websites for the people behind them. You never pick one: your search goes to the sources that fit.

Will it get my domain flagged?

Only deliverable addresses are sent by default. Sends are throttled and capped while your mailbox warms up, every email carries one-click unsubscribe, and openleads doctor grades SPF, DKIM and DMARC before your first send. A secondary sending domain is still a good idea.

Is cold email legal?

It depends on where you and your recipients are. OpenLeads keeps suppression lists and unsubscribe headers for you, and the responsible use guide covers CAN-SPAM, GDPR and CASL.

Do I need to know Python?

No. Install it with one command and type what you want, in the terminal, in the dashboard or through Claude. Python only matters if you want to add your own source.

Install it and describe your first customer.

Python 3.8 or newer. Read the docs