A desktop lead generation and enrichment system for discovering local businesses, scoring lead quality, storing data locally, and reviewing leads in one workflow. This project is still under development.
Atlas Engine: Building a Local-First Lead Generation and Lead Intelligence Platform
I am building Atlas Engine, a desktop lead generation and lead intelligence system designed to collect business data from public sources, enrich it with contact and website signals, score lead quality, and store everything in a local SQLite database for review and export. The project is focused on local businesses and includes a lead intelligence UI for approving, rejecting, and organizing prospects. It is still under development.
Why I built this
Most lead gen workflows either depend heavily on paid APIs or spread data across too many tools. I wanted something local-first, flexible, and built around a simple workflow: discover a business, enrich it, score it, review it, and export it. Atlas Engine is designed to do exactly that while keeping the data on the device.
What Atlas Engine does
The current root version of Atlas Engine includes a PySide6 desktop UI, a lead discovery engine, public web discovery through Google HTML, DuckDuckGo, and JustDial category pages, website crawling and extraction, lead scoring, SQLite storage, and a browser extension that sends signals to a local bridge.
In simple terms, the app works like this:
- You enter a niche, city, or area.
- The query generator expands search phrases and sources.
- The discovery engine collects candidate URLs from public search and directories.
- The scraper extracts page data, JSON-LD, and contact patterns.
- The scoring system classifies leads into Hot, Warm, or Cold.
- The data is stored locally in SQLite.
- You review, approve, reject, add notes, and export results.
Core features
Atlas Engine currently includes multi-source discovery, contact enrichment, lead scoring, lead management tools, website health analysis, local-first storage, browser page signals, and CSV export. The app also includes review actions such as approve/reject, quick actions, notes, and open-source-page shortcuts.
A few things I especially wanted in this build:
- Local-first data handling so the database stays on the device.
- Lead intelligence UI for fast decision-making.
- Browser extension signals that help identify scrapable business pages.
- Scoring and filtering to separate better leads from noisy ones.
- Export support for CRM or pipeline use.
Tech stack
Atlas Engine is built with Python 3.13, PySide6, and SQLite. The repo also includes a Chrome extension and a local HTTP bridge on 127.0.0.1:8765 for page-signal communication.
Architecture snapshot
The system is structured around a desktop UI, a scrape worker, public web sources, extraction modules, scoring logic, and a local database. The browser extension connects through a bridge and pushes page signals into the same workflow.
Versions and evolution
This repository also contains multiple older versions that show the evolution of the project. The earlier version was more API-driven, using Google Places API, Google CSE API, and OpenStreetMap sources. The newer current version is more scraping-first and adds browser signals plus website health checks. There is also a Streamlit-based version in the repo history.
Challenges and trade-offs
This kind of system is powerful, but it comes with real trade-offs. The README notes scraping fragility, anti-bot risk, lack of proxy rotation, local-only storage, a weak default login that should be changed, a missing popup.js for the extension popup, limited compliance tooling, and no automated tests yet. Those are exactly the kinds of issues that matter in a real lead gen workflow.
That is also why I like building it iteratively: every layer can be improved without losing the core workflow.
Who this is for
Atlas Engine is aimed at sales teams, local marketing agencies, business development reps, freelancers, growth teams, and researchers who need a practical way to build and review local business datasets.
Status
Atlas Engine is not presented as a finished commercial product yet. It is still under development, but the current version already shows a strong foundation for lead discovery, enrichment, scoring, and review.
Final thoughts
Atlas Engine is my attempt to combine lead generation, lead intelligence, local business enrichment, and a clean desktop workflow into one system. The goal is simple: make prospecting more structured, more local-first, and easier to control.
Try it now : https://github.com/BOSS294/Atlas-Engine
Built by: Mayank Chawdhari aka BOSS294
Company: Privonix Technologies — https://www.privonix.in/
Top comments (0)