Hi, I'm Rishindra Mateti, the developer and creator of this product.

Real LinkedIn outreach takes real research. You need to know who you're reaching out to, what they work on, what you have in common, and why you should start a conversation.

I turned that research process into an automation pipeline, so you can spend less time researching and more time deciding who to reach out to and what to actually say.

My first completely independent SaaS product, built around one core idea.

Don't automate the relationship. Automate the research behind it.

< 15 minOne-time setup
15+ minSaved per profile
Sign InCreate free account
Star on GitHub

What happens before you hit send.

We don't automate the relationship. We automate the research behind it. Here's exactly what networKING.agent looks for.

01

Research the person

Go beyond the name and job title. Understand their role, background, interests, and what they're actually working on.

02

Understand the company

Get the company context that matters: what they do, where the person fits, and what could be relevant to the conversation.

03

Find the connection

Surface real common ground: shared interests, overlapping experience, projects, or an angle worth opening with.

04

Know why you're reaching out

Turn the research into a clear reason to start the conversation, not just a collection of facts.

05

Keep your voice

Teach it how you write once. Drafts use your style and context instead of sounding like another AI template.

06

You hit send

No LinkedIn bot. No bulk blasting. You review the draft, edit it, and decide if and when it's sent.

AI researches. You decide.

How it flows.

Every profile moves through this precise sequence before a draft reaches your dashboard.

LinkedIn Profile
Research
Person & Company
Connection Points
Outreach Context
Your Decision
Your Message

Under the hood.

Multi-agent research pipeline

Profile research runs through specialized agent stages, built with Python, FastAPI, and Google Gemini.

Async worker orchestration

A queue of workers handles concurrency, retries, and automatic Gemini API key failover in the background.

Multi-tenant, secure by default

JWT authentication, per-user data isolation, and API keys encrypted at rest.

LinkedIn profile ingestion

Structural PDF parsing with duplicate contact detection, so a re-uploaded profile updates instead of duplicating.

Built on a real stack

Next.js, React, and Tailwind on the frontend. FastAPI, SQLAlchemy, and PostgreSQL on the backend.

Integrations

Slack and Telegram alerts, plus personalized email drafting for when a LinkedIn DM isn't enough.

What's next for v2.

  • Evidence-backed research with sources you can actually verify.
  • Relationship scoring to help decide who is worth reaching out to.
  • Conversation memory and smarter follow-ups.
  • Better company and person signals beyond basic profile information.
  • A networking graph connecting people, companies, interests, and relationships.

The goal isn't to automate networking.
It's to remove the most time-consuming part of it.