Get a demo

Address parsing & enrichment

Turn raw address strings from inbound emails and CSV uploads into clean, validated structured fields. Standardize street, city, state, zip, and country in one flow.

Get a demo

The prompt

I want to parse raw address strings from inbound emails and CSV uploads into clean structured fields. Can you build me a flow that extracts each address, splits it into street, city, state, zip, and country, standardizes the formatting, and outputs a validated CSV?

Just copy and paste the prompt into a new Parabola flow to get started.

What Parabola builds

A workflow with six steps you can edit:

1. Pull the inbound source. Monitor a shared inbox for new emails, or watch a folder for new CSV uploads. The flow ingests whichever source you have.

2. Extract the raw address. Pull the address string out of the email body, the subject line, or a column in the CSV. AI handles inconsistent placements and free-form text.

3. Split into structured fields. Street, unit, city, state or region, postal code, country. One column per field, regardless of how the source format ordered them.

4. Standardize the formatting. Uppercase state codes, ISO country codes, USPS-style street abbreviations, postal codes padded to the right length.

5. Validate. Flag missing fields, malformed postal codes, and country mismatches. The bad rows get tagged for review, not silently dropped.

6. Output the file. Clean CSV ready to load into your shipping system, CRM, HRIS, or fulfillment platform. Optional Slack alert when the count of flagged rows crosses a threshold.

Why teams stop doing this manually

Addresses come in dirty. A customer pastes a Google Maps result into a support ticket. A wholesale buyer emails a PO with the ship-to typed in the signature. A vendor uploads a CSV where the address is one long string in column G. Every downstream system needs the address split into clean fields, and a person has to do the splitting.

The manual version works for ten rows a day. It breaks the moment a CSV lands with 4,000 addresses, or a busy week of inbound emails buries the team. Mistakes here are not abstract. A misparsed zip routes a shipment to the wrong fulfillment center. A missing country tag fails a customs filing. A typo in a unit number gets a package delivered to the neighbor.

The work is high volume and rule-based. Extract, split, standardize, validate. That is exactly the kind of work that lives in a flow.

How it works

Step 1. Paste the prompt.

Open Parabola, paste the prompt in section 2, and let it ask follow-up questions about your inbound sources and the destination format your downstream system needs.

Step 2. Connect your data.

Email inbox, shared folder, or CSV upload. Plus the destination system if you want the clean file pushed automatically.

Step 3. Run it on a schedule.

Every fifteen minutes, hourly, or on each new inbound. The flow handles the parse, the standardization, and the validation. The team only looks at the flagged exceptions.

Related prompts

FAQ

Can the flow handle addresses outside the US?

Yes. The standardization step applies regional rules. Country codes are normalized to ISO 3166-1 alpha-2. Postal codes get country-specific validation.

What happens to addresses that the AI cannot parse cleanly?

They are tagged as flagged and routed to the exception list. The clean rows still flow through. A person reviews only the rows that need a judgment call.

Can the flow read addresses out of PDFs?

Yes. Add a PDF parsing step at the top of the flow. The AI extracts the text, finds the address pattern, and feeds it into the same standardization downstream.

Does this work for inbound order forms emailed by wholesale buyers?

Yes. The flow watches the inbox, pulls the address out of the body or attachment, and outputs a clean row. Common pattern for wholesale ops teams who receive POs by email.

How is this different from a standalone address validation API?

Validation APIs check whether an address is real. This flow extracts the address from messy inbound sources first, then validates, then standardizes for the destination system. The API can be a step inside the flow.
Clean addresses, every inbound.
Paste the prompt, point it at your inbox or upload folder, and let the parsing and validation run automatically.
Start for free