Built by John Peslar

We built the LinkedIn tool we couldn't find. Then we ran our own pipeline on it.

Zevari is the hosted LinkedIn MCP for Claude Code or Codex - the execution layer that lets your agents research, write in your voice, and stage every send for your approval. It works over MCP for Claude Code and Codex, or our REST API from your own code. It exists because the founder needed it to work, every day, on his own account. Not a demo. Not a deck. His actual pipeline.

The one-liner

Run LinkedIn outbound as code - and if you don't run Claude Code or Codex, we run it for you.

Meet the founder

John Peslar, GTM engineer

John Peslar is a solo founder and GTM engineer. He didn't come to LinkedIn outbound as a marketer - he came as a builder who got tired of watching Claude research a perfect prospect and then sit there, handicapped, because it couldn't actually touch LinkedIn. So he built the bridge. Zevari is the result.

Before Zevari, John grew his own LinkedIn from 1,000 to 50,000+ followers in 8 months - no team, no agency, just AI automations he wrote himself. That account has done 10M+ impressions. It feeds a newsletter with 40,000+ readers and Agent J, a 1,600-member community where founders and operators learn to deploy AI automation at scale.

Here is the part that matters: John runs his own outbound on Zevari every single day. The signal-based targeting, the Voice DNA trained on his sent messages, the approval gates, the warm-by-default sequencing, the daily digest - he uses all of it before you do. When something is clunky, he feels it first. When a connection ceiling or a pacing rule needs to hold, it holds because his own account is on the line.

That is the whole credibility claim, and it is the only one that counts in this category: an operator built this, an operator runs it, and an operator refuses to ship anything that would get his own account banned.

1,000 -> 50,000+LinkedIn followers in 8 months
10M+impressions on his own account
40,000+newsletter readers
1,600+members in the Agent J community

Shipped products

Four products that people pay for

Zevari is built and operated by John under LaunchpadFast, the parent company behind his earlier products. A small, operator-led team - which is the point.

LeadPanther

Inbound lead capture.

GetDeals

Outbound automation.

GetInterviews

AI job-search.

Zevari

The LinkedIn execution layer for Claude Code or Codex.

Why an operator-built tool is different

We optimize for the founder's own account surviving

Most LinkedIn automation gets built by people who do not have to live with the consequences. They optimize for a demo, bolt on browser cookies because it ships faster, and let the volume run because churn is somebody else's problem. That single constraint - the founder's account surviving, every day, for years - produced every real decision in the product.

Session-based connection, no browser cookies

Cookie-based tools are the ones getting accounts flagged. We never went near them, because John's account is the test account. Zevari connects through a managed session, not a scraped cookie injected into a headless browser.

Every write is staged for human approval

Message, connection request, comment, post - nothing leaves until a human clicks approve. That is not a feature we added. It is how John runs his own desk, every day.

Real ceilings, real pacing

Weekly connection ceilings (Free 40, Premium 150, Sales Navigator 200), working hours, behavioral pacing, duplicate checks, and burst caps. The numbers are published on the safety page because we actually enforce them.

A year of refinement, zero ban incidents

Not a marketing line - the operating record of the account this was built on. The model was refined over a year of the founder running his own outbound on it.

Safety is receipts, not adjectives

Anyone can write "safe" on a homepage. We publish the actual mechanics - staged approvals, no browser cookies, enforced weekly ceilings, behavioral pacing - and the actual track record of a year of real use with zero ban incidents.

Two ways to run the same engine

The engine John runs on himself is the same one you get. The only thing that changes is whose hands are on the keyboard.

Connect to Claude Code

Run the same engine the founder runs. Get the hosted LinkedIn MCP and the full safety model, self-serve. It works over MCP for Claude Code and Codex, or our REST API from your own code.

Connect to Claude Code

We run it for you

Don't run Claude Code or Codex? The same engine John runs on himself is available as a managed service. We set it up, run it on our infrastructure, and send you every message to approve from Slack.

We run it for you

Zevari - the LinkedIn execution layer for Claude Code or Codex. An operator built it, an operator runs it.