Address Normalization That Starts in the Email, Not the Spreadsheet

Parseur is a document parsing tool that reads the address out of the email or the PDF it arrived in, splits it into street, city, state, zip and country, and attaches latitude, longitude and a Google Maps link. All of that happens in one pass, while the document is processed, with no separate geocoding API and no address cleaning tool bolted on afterwards.

Key Takeaways

Addresses come out of the document already split into street, city, state, zip and country. No spreadsheet column required first.
Latitude, longitude and a map link ride along on every address, so a driver gets a pin instead of a street name.
Extraction, normalization and geocoding all run inside your document plan. No Maps API key to rotate, no per-lookup meter, no second vendor through procurement.
What Parseur will not do is certify USPS deliverability. If you need that, normalize here and hand the finished fields to a validation service.

Every address tool starts one step too late

Search for address normalization and every result assumes the same thing: your addresses already sit in a spreadsheet column, waiting to be cleaned. Smarty, Melissa, Loqate, Geocodio and the Google Maps Address Validation API all do excellent work from that starting line.

Nobody's addresses start there. They arrive inside a Zillow lead notification, a Shopify order confirmation, a delivery manifest PDF, a scanned work order. Somebody has to get them out first, and that somebody usually has a spreadsheet open on one monitor and an inbox on the other.

Parseur starts one step earlier. Hand it the document. It finds the address wherever it sits and returns the normalized components with coordinates attached. The cleanup step you were about to go shopping for is already done.

One house, three spellings, three customers

"123 Main St," "123 Main Street" and "Main St #123" are one building. Most systems file them as three. That is how a database ends up with duplicate records, failed deliveries and location data nobody trusts enough to route a driver by.

Parseur does not capture the address as a raw string. It normalizes the components into a standard shape, then enriches them with latitude, longitude and a map link. Readable was never the goal. Usable by the next system down the line was.

We see the same shape across the use cases that bring teams here, whether that is food delivery, real estate lead extraction, mortgage leads or e-commerce: contact details and an address, buried in an email, needed somewhere else within the hour.

Somebody is retyping these right now

Take the coffee shop owner taking orders online. Every order, he copies the delivery address out of the email by hand and passes it to a driver. Now put a dispatch desk in his chair, several hundred order emails and PDF work orders a day, and that same manual step running from open to close. Real estate agents run the loop too, on leads from Zillow, Realtor and Trulia: buyer contact details, plus the location of the property they asked about, street, zip and country.

Senders type an address however they feel like typing it, especially when no form ever checked the entry. A hundred emails a day eats a chunk of somebody's morning. A thousand is somebody's entire job, and it is the part of the job they resent most, particularly when every one of those addresses still has to land in a CRM such as Realvolve, Surefire CRM or an ERP.

You do not need a separate geocoding API

Most teams assume geocoding means wiring up a second service. Google Maps Platform, Smarty, Geocodio, Melissa, Loqate: all capable, all billed by usage, all wanting an API key, a rate limit to watch and a line item to defend at renewal. Google Maps Platform prices its Geocoding and Address Validation APIs per request, which is fine right up until the volume grows.

In Parseur, normalization and coordinates are part of document processing rather than a service you attach to it. They come with your document plan. No key to rotate, nothing metering in the background, no second vendor to put through procurement.

What Parseur does not do is postal validation. It will not certify that a mailpiece is deliverable to a USPS standard, because that is a different job with a different certification behind it. If deliverability certification is a requirement, normalize and geolocate with Parseur, then hand the clean components to a validation service. Routing and CRM matching only ever needed the components.

One email in, nine fields out

Forward a Zillow lead email and the property location comes back split, standardized and pinned to a coordinate:

Parseur field	Parsed data	Common name for it
PropertyAddress.normalized	12345 Random Rd, El Mirage, AZ 85335, USA	full standardized address
PropertyAddress.address1	12345 Random Rd	street_1
PropertyAddress.city	El Mirage	city
PropertyAddress.zip	85335	postal_code
PropertyAddress.state_code	AZ	state
PropertyAddress.country_code	US	country
PropertyAddress.lat	33.1234567	latitude
PropertyAddress.lng	-112.1234567	longitude
PropertyAddress.map	link to Google Maps	map link

The third column is the one that matters to whoever wires this into your stack. Whatever your CRM, ERP or routing tool calls these fields, the mapping is one to one, so connecting them is a field match instead of a parsing project.

Where this earns its keep

Real estate platforms

Leads from Zillow and Realtor.com arrive as unstructured email. Forward them to Parseur and the property address comes back geolocated, so agents get coordinates pushed to their phone or mapping tool through Zapier, instead of a street name to look up between viewings.

Last-mile delivery and field services

Order confirmations, dispatch requests and PDF work orders all carry whatever the customer typed into the box, and platforms like Shopify pass it through untouched. Parseur standardizes the address before it reaches the routing system. That is the difference between a planned route and a driver calling the office from the wrong street.

Directories and lists

Bulk sources such as business directory listings run the whole address together on one line, no commas, no line breaks, nothing to anchor to. Same treatment, same nine fields out.

Try out our powerful document processing tool for free.

Set it once, forget the format

Create a Parseur mailbox and forward the document to it. No template to build, no fields to highlight by hand.

Zillow Tour Request email sample

Tell Parseur which fields you want, in plain English, and the AI parsing engine finds them: Text AI for email bodies and text documents, Vision AI for PDFs, scans and images. Set the field format to address and Parseur normalizes and geolocates it automatically on every document that follows.

Set the field format to address

Field formats are how data normalization works across Parseur, on dates and numbers as much as on addresses. Set it once at the mailbox level and everything that arrives after lands in the same shape. Where a wrong address is expensive, switch on the optional human validation step and somebody eyeballs the parsed fields before anything exports.

From there the data goes where you need it: a CRM, a spreadsheet, a route planner, an API. What stops is somebody retyping an address a machine could have read the first time.

Last updated on September 8th, 2026

Ready to automate your
document data extraction?

Start free in minutes and see how Parseur fits into your workflow.

No model training required

Automates data entry from any document

Scales from point-and-click to API

Frequently Asked Questions

The questions teams ask before they let software touch their address data, starting with the one that decides the budget: how many tools does this actually take?

Address normalization is the process of turning a raw, inconsistently written address into a consistent set of labelled fields: street, city, state, postal code and country. "123 Main St", "123 Main Street" and "Main St #123" all normalize to the same record, which is what lets systems match, deduplicate and route on an address instead of on a string of text. Parseur normalizes addresses as it extracts them from the document, and adds latitude, longitude and a map link at the same time.

No. Parseur is not a CASS-certified address validation vendor and does not certify mail deliverability. What it does is read the address out of the document, split it into standard fields and geolocate it using Google Maps-backed data. Teams that need USPS deliverability certification on top of that pair Parseur with a validation service. For most workflows, the normalized, geocoded address is already what the CRM or the routing system was waiting for.

Email bodies, PDF attachments, scanned documents, images and spreadsheets. The Text AI engine handles emails and text documents, the Vision AI engine handles PDFs, scans and images, and the address comes back in the same normalized fields either way. This is the step address cleaning tools skip, because they assume the address is already sitting in a database column.

Parseur normalizes and geolocates addresses worldwide using Google Maps-backed data, so a French street address and a Japanese one come back structured exactly like a US one. Country codes are returned alongside the rest of the components.

That is the whole point of parsing by meaning instead of by position. A supplier can redesign its order confirmation email overnight and keep producing the same fields, because there is no per-sender template to rebuild when a layout changes.

Yes. Coordinates are ordinary fields, so they export to Google Sheets through the native integration, or anywhere else through Zapier, Make, n8n or a webhook.

Normalization reshapes an address into standard fields. Validation checks against a postal authority database, such as the USPS, that the address exists and can receive mail. Different jobs, different questions. Parseur handles normalization and geolocation on addresses it pulls out of your documents. If your business depends on postal deliverability certification, run the normalized output through a dedicated validation service afterwards.

Not for extraction, normalization or geocoding. Those tools take an address you already have and clean it up, which means you still need something to get the address out of the email or the PDF first. Parseur does both steps, and address normalization and coordinates are included in your document plan rather than billed per lookup. You would add one of those services only for postal validation, which Parseur does not do.

That is the normal case, not the hard case. Parseur reads the address by meaning rather than by position, so a run-on string like "12345 Random Rd El Mirage AZ 85335" is split into street, city, state and zip without you marking the boundaries. Nothing depends on commas, line breaks or the layout of the document it arrived in.

Parseur resolves the best possible match from the surrounding context and the available location data, and returns the most likely result. Where a wrong address is expensive, turn on the optional human validation step and review parsed addresses before they are exported.

It is what makes deduplication possible. Two records reading "123 Main St" and "123 Main Street" look like two customers to any system comparing raw text, and like one customer once both are normalized to the same components. Normalizing at intake, before the record is created, costs less than cleaning the database later.

Every component, including the map link and the GPS coordinates, can be pushed to CRMs, ERPs, dispatch tools and route planners through the native integrations or a webhook. Field agents get coordinates rather than a street name somebody still has to look up.