Research and ICP scoring
Prospects scored 1 to 5 against your ICP with the why attached, so the agent skips the 2-out-of-5s and learns when you correct it.
Zevari vs Meet Alfred
Meet Alfred runs LinkedIn from your browser. Zevari runs it from inside Claude.
Meet Alfred is a mature browser-based LinkedIn automation tool - a Chrome extension or cloud session that logs into LinkedIn on your behalf and fires templated sequences. That is the model the entire first generation of LinkedIn tools is built on.
Zevari is the hosted LinkedIn MCP for Claude. Your agent researches, scores prospects against your ICP, writes in your voice, and stages every send for your approval - connecting through a session with no browser cookies. Run LinkedIn outbound as code, and if you don't run Claude Code or Codex, we run it for you.
The frame
We share a buyer, not a category
If you run Meet Alfred, you already believe LinkedIn is your best channel and you already automate it. We agree. The question this page answers is narrower: should the work happen in a browser extension that fires templated sequences, or inside the AI you already think with?
One customer put it plainly on a call: "maybe I don't need Meet Alfred anymore because Zevari is doing all the stuff." That's the whole pitch - not "Alfred is bad," but "the stack you're paying for - the scraper, the sequencer, the enrichment, the second AI tool you bolted on - collapses into one agent that lives in Claude."
Lookalike from a known-good client
Hand it one customer or a single profile URL and it finds more like them - no filter-stacking, no second scraper.
Inbox Radar
Replies classified by intent, on-brand responses staged for your approval, so you stop scrolling for the one reply that matters.
Campaigns and follow-ups
Built and advanced by your agent on schedule, with email as the follow-up leg to LinkedIn non-responders - not a separate sequencer.
Where the two diverge
Three reasons buyers switch
Three things separate a Claude-native agent from a browser automator. They're the three reasons people move.
Targeting is signal-based, not list-based
Meet Alfred runs the list you hand it. Zevari scans for people who posted about your topic in the last 30 days, scores each prospect 1 to 5 against your ICP with the reasons written out, and reaches only the high-intent ones. Stop blasting a static export; start reaching people who are in-market this week.
The messages sound like you
Meet Alfred sends the template you wrote, with variables swapped in. Zevari's Voice DNA is trained on your own sent messages, so the drafts read like you wrote them - not like merge-tag mail. The AI version of you, going out to do its thing.
The whole thing runs from one brain
With Alfred you assemble a stack - the automator, a scraper, sometimes a separate AI writer, a CRM connector. With Zevari your agent does the research, scoring, warming, campaigns, inbox triage, and email follow-up to non-responders. One console, one approval surface, one place the strategy lives.
Where it lives
Inside Claude, not a browser tab
Meet Alfred is an extension you install and a UI you log into. Zevari is a hosted MCP server your agent connects to - 60+ tools for search, signals, campaigns, inbox, and content. Reach Zevari over MCP for Claude Code and Codex, or our REST API from your own code.
Connect in one line
claude mcp add zevari https://mcp.zevari.ai/mcpThat means your GTM brain and your execution layer are the same agent. Claude can already research a prospect deeply; with Zevari it can finally act on what it found - connect, message, comment, post - without you stitching a scraper plus a sequencer plus an AI writer plus a CRM connector into one fragile stack. And because Zevari holds hosted state, your campaigns keep advancing and your follow-ups keep firing on schedule - the persistent scheduling Claude Code or Codex can't do on its own.
Safety
"I don't want to be banned." That's why browser-based tools lose the buyer.
When a tool logs into LinkedIn through a browser session and fires sequences on a schedule, LinkedIn sees the logs - and accounts get flagged. Zevari's answer isn't a louder adjective; it's the mechanics. A year of refinement. Zero ban incidents. We publish the actual numbers because nobody else does.
Read the full safety modelSession-based connection. No browser cookies, no extension logging into LinkedIn on your behalf.
Every write action - message, connection request, comment, post - is staged for your approval before it touches your account.
Hard daily connection ceilings: Free 40 / Premium 80 / Sales Navigator 150 per week.
Working hours, behavioral pacing, duplicate checks, and burst caps - so you can't overshoot by accident.
The comparison, straight
An honest table
Meet Alfred does real work and a lot of teams are happy with it. What it can't be is Claude with hands on LinkedIn - that's a different architecture, not a feature toggle.
Meet Alfred
Zevari
Where it runs
Meet Alfred
Chrome extension / cloud browser session
Zevari
Inside Claude Code or Codex (or any MCP client)
Connection model
Meet Alfred
Logs into LinkedIn for you (browser-based)
Zevari
Session-based connection, no browser cookies
Targeting
Meet Alfred
Lists and saved searches you build
Zevari
Signal-based: posted-in-30-days, ICP scoring 1-5 with reasons
Message writing
Meet Alfred
Templates with variables
Zevari
Voice DNA - trained on your sent messages
Sends
Meet Alfred
Sequences fire automatically
Zevari
Every write staged for human approval (message, connection, comment, post)
Warming
Meet Alfred
Auto-endorse / view add-ons
Zevari
Warm-by-default: comments, reactions, profile views before the ask
Inbox
Meet Alfred
Reply detection, basic
Zevari
Inbox Radar - replies classified by intent, drafts staged
Safety posture
Meet Alfred
Weekly limits, randomization
Zevari
Daily ceilings, working hours, behavioral pacing, duplicate checks, burst caps; a year of refinement, zero ban incidents
Done-for-you option
Meet Alfred
No
Zevari
Yes - a managed tier runs the whole engine and you approve from Slack
Strategy lives in
Meet Alfred
The tool's UI
Zevari
Your Claude agent's context - your GTM brain
When to stay on Meet Alfred: if you want a self-contained browser tool, you don't work inside Claude Code or Codex, and templated sequences against lists you build are getting you the volume you need - it's a solid tool for that job. Switch to Zevari if you already run Claude (or want to), you're tired of stitching a scraper plus a sequencer plus an AI writer plus a CRM connector into one fragile stack, you want signal-based targeting and messages that sound like you, or "I don't want to be banned" is the sentence keeping you up.
FAQ
Questions buyers ask when they're switching
Is Zevari a Meet Alfred alternative?
Yes, for the same job - running LinkedIn outbound - but it is a different architecture. Meet Alfred is a browser-based automation tool that logs into LinkedIn for you and fires templated sequences. Zevari is a Claude-native agent that researches prospects, scores them against your ICP, writes in your voice, and stages every send for your approval, connecting through a session with no browser cookies. Many buyers consolidate their stack onto Zevari rather than run both.
Will Zevari get my LinkedIn account banned?
No incidents in a year of refinement. Zevari uses a session-based connection (no browser cookies), stages every write action for human approval, and enforces hard limits: daily connection ceilings (Free 40, Premium 80, Sales Navigator 150 per week), working hours, behavioral pacing, duplicate checks, and burst caps. The full mechanics are published on the safety page.
Do I need Meet Alfred and Zevari both?
Most people do not. Zevari does the research, ICP scoring, voice-matched message writing, campaigns, warming, inbox triage, and email follow-up to non-responders in one place. The common path is to consolidate the stack - one customer told us "maybe I don't need Meet Alfred anymore because Zevari is doing all the stuff."
Do I need LinkedIn Sales Navigator to switch?
No. You can start without it. Sales Navigator only raises your weekly connection limit (to 150 per week versus 80 on Premium and 40 on Free) - it is not required to run Zevari.
Can it find people who are posting about a topic and reach out to them?
Yes. That is signal-based targeting: Zevari scans for people who posted about your topic in the last 30 days, scores each one 1 to 5 against your ICP with written reasons, and reaches only the high-intent prospects - instead of running a static list export.
Does it actually send messages, or just draft them?
It drafts, and you approve. Every write action - message, connection request, comment, post - is staged for your sign-off before anything is sent. That human-in-the-loop gate is the safety model, not a setting you toggle off.
I'm not technical. Can I still switch off Meet Alfred?
Yes. If you don't run Claude Code or Codex, our managed tier runs the whole engine for you - setup, scheduled workflows, tuning - and you approve sends from a simple Slack digest in minutes a day. It is application-gated for buyers with an established offer and paying customers, so it replaces what you already spend on an appointment setter.
Consolidate the stack into one agent
Same approval-gated engine on both lanes. Reach Zevari over MCP for Claude Code and Codex, or our REST API from your own code.
You run Claude Code or Codex
Connect Zevari to your agent in 60 seconds - signal-based, ban-safe, in your voice, with human-in-the-loop approval gates baked in. Connect over MCP for Claude Code and Codex, or call our REST API from your own code.
Connect to Claude CodeWe run it for you
Don't run Claude Code or Codex - and already paying an appointment setter or AI SDR? We run the same engine - research, voice-matched messages, campaigns, replies - and you approve every send from Slack in minutes a day.
We run it for youZevari - the LinkedIn execution layer for Claude. Your agent sleeps; your pipeline doesn't.