Mailparser

Consume the structured data Mailparser extracts from inbound emails. Agents list mailboxes and their inbound addresses, retrieve parsed results over a date range, and look up individual parsed emails to route extracted fields downstream.

5 actions API key available

Mailparser website ↗ Platform Documentation ↗ Capability data verified 2026-07-28

A schedule starts the hourly sweep of the invoices mailbox

List Parsed Results pulls extracted fields since the last run: vendor, invoice number, total, due date

Get Parsed Email fetches the full record for each new result, including sender and subject

The agent matches vendor and invoice number against open purchase orders

Invoices with no matching PO, or totals outside the PO tolerance, route to the AP manager with the parsed fields and source email side by side

Matched invoices post to the accounting system as bills awaiting payment run

The AP channel gets the hourly digest: parsed, matched, and held counts

What This Integration Enables

An enormous amount of business still arrives as email: invoices, order confirmations, lead notifications, shipping updates. No integration roadmap eliminates that, because the senders are other companies with their own systems and no intention of changing them. Mailparser turns those emails into structured fields; this connector is how FlowRunner agents pick those fields up and put them to work. The pattern is simple: each Mailparser mailbox has an inbound address and parsing rules, and the agent's job is to sweep what those rules extract and route it downstream while the email is still current. - Pull extracted fields, invoice numbers, totals, due dates, order data, into workflows over a date range - Sync parsed email data into spreadsheets, CRMs, and accounting systems on a schedule - Look up individual parsed emails by id, with sender, subject, and received time for context - Enumerate mailboxes and their inbound addresses to audit the email-to-data pipeline itself, end to end The division of labor is clean: Mailparser owns the parsing rules and the inbound addresses, and FlowRunner owns everything after the parse, the scheduling, the matching, the routing, and the record of what happened to every extracted field. Teams tune extraction in one tool and orchestrate consequences in the other, and neither has to impersonate the other's job. Extraction is probabilistic in practice: templates change and rules drift. That is why the money-moving steps downstream of the parse run through FlowRunner's [human-in-the-loop](/concepts/human-in-the-loop) gates rather than straight to execution.

Without FlowRunner

Inbox as inbox Invoices, orders, and leads sit in a mailbox until a person opens each one and retypes it

Parsed data stranded Mailparser extracts the fields, then the structured data waits for someone to export it

Silent parsing drift A vendor changes their template, a field starts extracting wrong, and nobody notices for a quarter

With FlowRunner

Inbox as data source Agents sweep parsed results on a schedule and the extracted fields flow straight into the workflow

Every record lands somewhere Parsed invoices become bills, parsed leads become contacts, parsed orders become fulfillment tasks

Exceptions surface with evidence Records that fail matching arrive at a person with parsed fields and source metadata attached

Use Case Scenarios

The invoice inbox that files itself

Vendors email invoices to the Mailparser inbound address. Hourly, the agent runs List Parsed Results with a from date of the last sweep, appends each record to [Google Sheets](/integrations/google-sheets) as the AP intake log, and creates draft bills in [QuickBooks Online](/integrations/quickbooks). Anything that fails the PO match holds for the AP manager. Nobody retypes an invoice, and nobody pays one on a parser's word alone, which is the balance an AP process actually wants.

Inbound leads reach the CRM before they cool

Lead notification emails from directories and partner sites land in a dedicated mailbox with rules extracting name, email, company, and inquiry. The agent sweeps List Emails, pulls each record with Get Parsed Email, dedupes against [HubSpot](/integrations/hubspot) with a contact search, and creates the contact with the inquiry attached. The lead that used to wait in a shared inbox until Monday gets a same-hour follow-up task, and the dedupe step means the third notification about the same prospect enriches one record instead of creating three.

Order emails drive fulfillment, with a drift alarm

Marketplace order confirmations parse into order number, items, and shipping address. The agent fetches each with Get Parsed Email and posts the order to the fulfillment channel in [Slack](/integrations/slack). It also runs a nightly sanity pass: when the count of parsed results diverges sharply from the count of received emails in List Emails, or required fields start coming back empty, it flags the mailbox to operations. Parsing drift becomes a same-day alert instead of a quarter-end surprise, caught by the cheapest possible check: two counts that should match, compared nightly.

Human-in-Loop Highlight

Get Parsed Email returns whatever the mailbox's parsing rules extracted, and the dangerous day is the one after a vendor redesigns their invoice template: the rule still fires, the field still populates, and the total is now the order number. Every step downstream of that parse is real, a bill posted to the ledger, a payment queued, a contact created, so the FlowRunner agent treats parsed money fields as claims to verify, not facts to act on. Invoices that match an open PO within tolerance flow through. Anything else, no PO, total out of band, a vendor whose fields shifted, stops in front of the AP manager with the parsed values and the source email's sender and subject side by side. Approving a mismatched invoice takes a person seconds. Clawing back a payment sent on a misparsed total takes considerably longer.

Agent processes routinely

Detects exception requiring judgment

Clear match Continues automatically

Ambiguous Routes to human via email

Human decides

Agent resumes with decision

Agent Capabilities

5 actions

Mailboxes

List Mailboxes Lists all parsing inboxes in the account, each with its unique inbound email address and parsing rules. The starting point for any sweep and the audit view of the email-to-data pipeline.
Get Mailbox Retrieves one mailbox's configuration by id, including its inbound address, name, and metadata. Used when a workflow needs to verify where a data stream actually comes from.

Parsed Data

List Parsed Results Lists the structured field records a mailbox's rules have extracted, scoped by an optional date range and limit. The workhorse read that pulls processed data into workflows on a schedule.
Get Parsed Email Retrieves a single parsed email's extracted fields by id, with source metadata: sender, subject, received time. The record-level read behind matching, enrichment, and exception review.

Emails

List Emails Lists the raw emails a mailbox has received with sender, subject, and received time. Paired with parsed-result counts, it is also the drift detector: when received and parsed diverge, something changed upstream.

Frequently Asked Questions

What can FlowRunner do with Mailparser?

FlowRunner agents can run List Mailboxes, Get Mailbox, and List Parsed Results in Mailparser, plus 2 more actions.

Does connecting Mailparser to FlowRunner require OAuth?

No. Mailparser connects to FlowRunner with an API key, no OAuth flow required.

Can Mailparser trigger a FlowRunner workflow automatically?

Mailparser doesn't currently expose triggers in FlowRunner. It connects as an action step inside workflows started by another trigger.

Start building with Mailparser

$100 in credits. No card required. Connect in minutes.

Start Building Free Talk to us