Map lead JSON with Claude CodeTurn exported business data into a living map
Capture leads on Google Maps, export the JSON, and let Claude Code or ChatGPT plot every business on an interactive map. It needs no server, API key or coding.
AI Workflows··8 min read
Key takeaways
The Vonsel Chrome extension captures businesses from Google Maps and exports them as CSV, Excel, or JSON, the ideal format to feed an AI
Claude Code and ChatGPT can read that JSON and write a full interactive map for you, plotting each lead by its coordinates
The extension only collects the data. Your AI (and the Vonsel dashboard) turns raw rows into decisions and outreach
Yes, Claude Code can turn a lead JSON file into an interactive map. Export your Google Maps leads from the Vonsel extension as JSON, put the file in a folder, and ask for a single HTML page. Each business is plotted from its latitude and longitude, so you need no geocoding, server or API key.
The idea
Feed your AI with real business data
A list of leads in a spreadsheet is easy to ignore. The same leads plotted on a map, colored by rating and clustered by neighborhood, tell you where to spend your week. Building a custom map used to mean hiring a developer, and now it doesn't.
Every business the Vonsel extension captures already carries its latitude and longitude, name, rating, review count, phone, and website. Export that as JSON and you are holding a clean, structured file that an AI can read in one pass. The extension does the collecting. Your AI does the building.
This guide walks the practical flow: capture on Google Maps, export JSON, then prompt Claude Code or ChatGPT to turn it into a map you can open in your browser. If you want the whole "feed your AI" mindset first, start with feed your AI with business data.
Start by capturing the data
Install the free extension, scrape a city on Google Maps, and export the JSON in one click. Free download. No trial, no credit card.
export formats from the extension: CSV, Excel, and JSON
1
plain-language prompt to build an interactive map
0
lines of code you need to write yourself
Step 1
Capture and export the JSON
Open Google Maps, search for your niche and area (for example "dentists in Austin"), and let the extension capture the results into its side panel. When the list looks right, choose Export and pick JSON. For a refresher on the formats, see export leads as CSV, Excel, or JSON for AI.
JSON is the format to reach for here because it keeps the structure intact. Each lead is an object with named fields, so the AI does not have to guess which column is the latitude. A typical exported record looks like this:
Claude Code runs in your terminal and can read files, write files, and open them, which suits this job. Drop your leads.json in a fresh folder, start Claude Code there, and give it one clear instruction:
"Read leads.json. Build a single self-contained index.html that plots every business on an interactive map using its lat and lng. Color each marker by rating (green above 4.5, amber 4.0 to 4.5, red below 4.0) and show a popup with the name, rating, review count, and phone. Then open it in my browser."
Claude Code will pick a mapping library, wire up the markers, handle the popups, and write the whole page. Because it works in a real project folder, you can iterate by conversation: "cluster markers when zoomed out", "add a filter for businesses with no website", "export the filtered set back to JSON". Each request edits the same file. For deeper patterns, see build a lead dashboard with Claude Code and analyze lead lists with Claude Code.
If you would rather hand the AI a standards-based file, ask it to convert your JSON into GeoJSON first, a format most mapping tools read natively.
Step 3
Or do it in ChatGPT with no setup
If you prefer a browser tab to a terminal, ChatGPT works just as well for a one-off map. Upload the exported JSON directly into the chat, then ask for the same outcome: an HTML file that maps every lead by coordinates. ChatGPT returns the code (and, with its data tools, can even preview a static map), and you save the file and open it locally.
ChatGPT is fastest for a single map you want in five minutes. Claude Code is better when the map becomes a project you keep refining, or when you want it reading and rewriting files on your machine. Deciding which format to hand each one? JSON vs CSV for AI analysis settles it.
You want
Reach for
A quick one-off map from a single export
ChatGPT
An evolving dashboard you tweak over days
Claude Code
Batch scripts across many export files
Codex / local LLM
Offline or private data handling
Local LLM
For teams with strict privacy rules, a local model (run through Ollama, LM Studio, or similar) can produce the same map code without your lead file ever leaving the machine. It is slower and needs more prompting, but the JSON contract is identical, so the workflow does not change.
Get the JSON your AI needs
The extension exports clean, coordinate-rich JSON that any AI can map. Free download. No trial, no credit card.
Every captured business ships with lat and lng, so the AI never has to geocode addresses or guess locations. The map is accurate from the first render.
Structured fields, not loose text
Named JSON keys mean the AI maps rating to color and reviews to size without you explaining the schema, so you prompt less and it makes fewer mistakes.
No API keys or servers
The map renders locally without an API key. The output is a single HTML file you can email, host, or open offline.
Iterate in plain language
Filters, clusters, and color rules are one sentence away. You describe the territory view you want and the AI edits the code.
A map is a decision tool, not a decoration. Once you can see clusters of unhappy competitors, gaps with no coverage, or high-rating targets by street, your prospecting stops being a spreadsheet and starts being a route.
The bigger picture
When you want the map built for you
Building your own map with AI is flexible, and for many analysts it is the whole point. But not everyone wants to prompt a model every time they open a city. The Vonsel dashboard is for them.
Inside the mapped CRM, captured leads already appear on an interactive map with sales zones and routes. Reviews Intelligence surfaces competitor pain points, Email Intelligence drafts a tailored message per business, and value-added databases arrive contextualized instead of raw. The extension and your AI are the workshop. The dashboard is the finished machine.
Most people use both: the extension plus a custom AI map for exploration, and the dashboard when it is time to work leads at scale. See how the pieces connect in from search to CRM: the complete workflow, or browse the full AI Workflows series.
The extension hands you the raw material. Your AI turns it into a map you act on.
Ready to feed your AI real data?
Capture Google Maps businesses and export the JSON your AI needs to build maps, dashboards, and outreach. Free download. No trial, no credit card. Explore features or see plans.
Can Claude Code build a map from a lead JSON file?
Yes. Export your leads as JSON from the Vonsel extension, drop the file in a folder, and ask Claude Code to build an interactive map from it. It writes a small HTML page that plots every business by latitude and longitude, adds popups with name, rating, and phone, and opens in your browser. No server and no API key needed.
Do I need to know how to code to map my leads with AI?
No. Both ChatGPT and Claude Code accept plain-language prompts. You upload the exported JSON or CSV, describe the map you want, and the AI writes the code. Your job is to describe the outcome, not to write JavaScript.
Is exporting lead data from the Vonsel extension free?
Yes. The Vonsel Chrome extension is a free download with no trial and no credit card. You capture businesses from Google Maps and export them as CSV, Excel, or JSON, then feed that file to any AI tool you like.
Which fields from the lead JSON does the map use?
The name becomes the popup title, lat and lng set the marker position, rating sets the marker color, and reviews sets the marker size. Phone and website add detail to the popup. Because every captured business already carries its coordinates, the AI never has to geocode addresses.
Can I change the map after Claude Code builds it?
Yes. Claude Code works in a real project folder, so each request edits the same file. You can ask it to cluster markers when zoomed out, add a filter for businesses with no website, or export the filtered set back to JSON.
Can I share the finished map with someone else?
Yes. The output is a single HTML file that renders locally without an API key, so you can email it, host it, or open it offline.
What is GeoJSON and do I need it to build the map?
GeoJSON is a standard format for geographic data, defined in RFC 7946 by the IETF. You do not need it, because Claude Code can read the Vonsel lead JSON directly and plot it. It helps when you want to open the same leads in other mapping tools, and you can ask the AI to convert the file in one prompt.
Why does the map not need geocoding?
Geocoding is the process of turning an address or place name into latitude and longitude. Every business captured from Google Maps already carries its coordinates in the exported JSON, so the AI places each marker straight from the lat and lng fields. That removes a step and avoids addresses being matched to the wrong spot.
Sources
Claude Code overview (Claude Code Docs): describes Claude Code as an agentic tool that reads your codebase, edits files and runs commands, available in your terminal.
GeoJSON: the home page of the format that most mapping tools read natively.
