email-forward-parser-py

email-forward-parser extracts the original message details from forwarded email content. It can work with plain forwarded bodies or full .eml messages, including common MIME encodings and attached .eml files.

The package is useful when you receive emails forwarded by users or mailboxes and need to recover the original sender, recipients, subject, sent date, and body.

Features

• Detects forwarded messages from subject prefixes such as Fwd: and FW:.
• Parses forwarded bodies from Gmail, Apple Mail, Outlook/Office 365, Thunderbird, Yahoo Mail, HubSpot, Missive, IONOS, and several localized clients.
• Extracts original metadata: From, To, Cc, Subject, Date, and body.
• Parses full .eml messages with MIME decoding.
• Handles quoted-printable/base64 text bodies and attached message/rfc822 or .eml attachments.
• Rebuilds an .eml string for the original message when a forwarded email is detected.
• Ships with a 100% statement and branch coverage gate.

Installation

python -m pip install email-forward-parser

Python 3.14 or newer is required.

The distribution name is email-forward-parser, but the import package is emailforwardparser:

from emailforwardparser.forward_parser import get_forwarded_metadata
from emailforwardparser.client import EmailParserClient

Quick Start

Use get_forwarded_metadata() when you already have the email body text.

from emailforwardparser.forward_parser import get_forwarded_metadata

body = """Hi team

---------- Forwarded message ---------
From: Jane Doe <jane@example.com>
Date: Mon, 1 Jan 2024 at 12:00 PM
Subject: Original subject
To: Bob <bob@example.com>
Cc: Copy <copy@example.com>

Original body.
"""

result = get_forwarded_metadata(body, "Fwd: Original subject")

print(result.forwarded)             # True
print(result.message)               # Hi team
print(result.email.subject)         # Original subject
print(result.email.from_.name)      # Jane Doe
print(result.email.from_.address)   # jane@example.com
print(result.email.to[0].address)   # bob@example.com
print(result.email.cc[0].address)   # copy@example.com
print(result.email.body)            # Original body.

Use EmailParserClient when you have a full .eml message.

from emailforwardparser.client import EmailParserClient

client = EmailParserClient()

with open("forwarded.eml", encoding="utf8") as file:
    raw_eml = file.read()

metadata = client.get_original_metadata(raw_eml)

if metadata.forwarded:
    print(metadata.email.from_.address)
    print(metadata.email.subject)
    print(metadata.email.body)

API

get_forwarded_metadata(body, subject=None)

Parses a forwarded body string and returns a ForwardMetadata dataclass.

from emailforwardparser.forward_parser import get_forwarded_metadata

result = get_forwarded_metadata(body, subject)

subject is optional. Passing it helps the parser detect forwards that have a forwarded subject prefix but no separator line in the body.

ForwardMetadata

@dataclass
class ForwardMetadata:
    forwarded: bool
    message: str
    email: OriginalMetadata

• forwarded: True when forwarded content was detected.
• message: the note written by the person who forwarded the email, before the forwarded block.
• email: metadata extracted from the original email.

OriginalMetadata

@dataclass
class OriginalMetadata:
    date: str
    subject: str
    body: str
    from_: MailboxResult
    to: list[MailboxResult]
    cc: list[MailboxResult]

The sender field is named from_ because from is a Python keyword.

MailboxResult

@dataclass
class MailboxResult:
    name: str
    address: str

When a mailbox line has no valid email address, address is empty and the raw value is kept in name.

EmailParserClient

EmailParserClient is the high-level wrapper for full .eml messages.

from emailforwardparser.client import EmailParserClient

client = EmailParserClient()

Methods:

• get_original_metadata(email: str) -> ForwardMetadata Parses a raw .eml string and returns metadata for the original email. If the message is not forwarded, it returns metadata for the message itself.
• get_original_metadata_from_file(file_path: str) -> ForwardMetadata Reads a .eml file and returns the same metadata object.
• get_original_eml(email: str) -> dict Returns a dictionary with:
  • forward: whether the input was detected as forwarded.
  • Send-To: the first address from the wrapper message's From header.
  • eml: a raw .eml string for the original message when forwarded, or the input/attached message when not forwarded.
• get_original_eml_from_file(file_path: str) -> dict Reads a .eml file and returns the same dictionary.

Examples

Parse an Apple Mail Forward Without a Subject

from emailforwardparser.forward_parser import get_forwarded_metadata

body = """Personal note

Begin forwarded message:

From: Jane Doe <jane@example.com>
Date: Mon, 1 Jan 2024 at 12:00 PM
Subject: Original subject
To: Bob <bob@example.com>

Original body.
"""

result = get_forwarded_metadata(body)

assert result.forwarded is True
assert result.message == "Personal note"
assert result.email.subject == "Original subject"
assert result.email.from_.address == "jane@example.com"

Parse a Quoted-Printable .eml

from emailforwardparser.client import EmailParserClient

raw_eml = """From: Forwarder <forwarder@example.com>
To: parser@example.com
Subject: Fwd: Original subject
MIME-Version: 1.0
Content-Type: text/plain; charset="utf-8"
Content-Transfer-Encoding: quoted-printable

Hi=0A=0A---------- Forwarded message ---------=0AFrom: Jane Doe <jane@example.com>=
=0ADate: Mon, 1 Jan 2024 at 12:00 PM=0ASubject: Original subject=0ATo: Bob =
<bob@example.com>=0A=0AOriginal caf=C3=A9 body.=0A
"""

metadata = EmailParserClient().get_original_metadata(raw_eml)

assert metadata.forwarded is True
assert metadata.email.body == "Original café body."

Rebuild the Original .eml

from email.message import EmailMessage

from emailforwardparser.client import EmailParserClient

message = EmailMessage()
message["From"] = "Forwarder <forwarder@example.com>"
message["To"] = "parser@example.com"
message["Subject"] = "Fwd: Original subject"
message.set_content("""Hi

---------- Forwarded message ---------
From: Jane Doe <jane@example.com>
Date: Mon, 1 Jan 2024 at 12:00 PM
Subject: Original subject
To: Bob <bob@example.com>

Original body.
""")

data = EmailParserClient().get_original_eml(message.as_string())

assert data["forward"] is True
print(data["eml"])

Read From a File

from emailforwardparser.client import EmailParserClient

client = EmailParserClient()

metadata = client.get_original_metadata_from_file("forwarded.eml")
data = client.get_original_eml_from_file("forwarded.eml")

print(metadata.email.subject)
print(data["eml"])

Non-Forwarded Messages

If a message is not forwarded, forwarded is False. With the client wrapper, the returned email metadata describes the message itself.

from emailforwardparser.client import EmailParserClient

metadata = EmailParserClient().get_original_metadata(raw_eml)

if not metadata.forwarded:
    print("This was not a forwarded message")
    print(metadata.email.subject)

Behavior Notes

• Parsing is regex-based. It is designed for common forwarded-message formats, not every possible custom email template.
• Missing fields are returned as empty strings or empty lists.
• Multiple To and Cc recipients are returned in order.
• The parser normalizes common email encodings and non-breaking spaces.
• The client prefers the first non-attachment text/plain body part when reading multipart .eml messages.
• Attached .eml messages are returned directly when present.

Development

Install the development and test tools:

python -m pip install -e ".[dev,test]"

Run the full local quality gate:

python -m isort --check-only emailforwardparser tests
python -m black --check emailforwardparser tests
python -m ruff check emailforwardparser tests
python -m mypy
python -m pytest

pytest is configured to require 100% statement and branch coverage.

Release

Build and verify the distribution:

python -m pip install --upgrade build twine
python -m pytest
rm -rf dist build
python -m build
python -m twine check dist/*

Upload to PyPI:

python -m twine upload dist/*

License and Credits

This project is distributed under the Apache License, Version 2.0. See LICENSE for the full license text.

The project was originally created by Garrett Marking. This fork/package includes additional maintenance, packaging, typing, test coverage, and parser/client fixes by Felipe Hertzer.

Apache 2.0 allows redistribution and modification, including publishing to PyPI, as long as the license conditions are followed. In practice for this package:

• Keep the Apache 2.0 license text with redistributions.
• Keep existing attribution notices from the source distribution.
• Keep the NOTICE file with source and binary distributions.
• Make it clear when files have been modified.

For the canonical license terms, refer to the Apache Software Foundation's Apache License 2.0 page.