Outlook Email Message (.msg) parser

A parser for Microsoft Outlook .msg files (OLE Compound Document format). Extracts message metadata, body content, recipients, attachments, and transport headers as specified in [MS-OXMSG] and [MS-OXPROPS].

Usage

Add this to your Cargo.toml:

[dependencies]
msg_parser = "0.3"

Quick Start

use msg_parser::Outlook;

fn main() {
    let outlook = Outlook::from_path("email.msg").unwrap();

    println!("From: {} <{}>", outlook.sender.name, outlook.sender.email);
    println!("Subject: {}", outlook.subject);

    for person in &outlook.to {
        println!("To: {} <{}>", person.name, person.email);
    }
    for person in &outlook.cc {
        println!("CC: {} <{}>", person.name, person.email);
    }
}

Parsing from different sources

use msg_parser::Outlook;

// From a file path
let outlook = Outlook::from_path("email.msg").unwrap();

// From a byte slice (accepts &[u8], Vec<u8>, or anything AsRef<[u8]>)
let bytes = std::fs::read("email.msg").unwrap();
let outlook = Outlook::from_slice(&bytes).unwrap();

// Passing a Vec<u8> directly also works
let outlook = Outlook::from_slice(bytes).unwrap();

// From any std::io::Read source (file, stdin, network, etc.)
let file = std::fs::File::open("email.msg").unwrap();
let outlook = Outlook::from_reader(file).unwrap();

Display formatting

Outlook, Person, and Attachment all implement Display for human-readable output:

let outlook = Outlook::from_path("email.msg").unwrap();
// Prints a summary: From, Subject, To, CC, BCC, Date, Attachments
println!("{}", outlook);

Saving attachments

let outlook = Outlook::from_path("email.msg").unwrap();

for attach in &outlook.attachments {
    let name = if attach.long_file_name.is_empty() {
        &attach.file_name      // 8.3 short name fallback
    } else {
        &attach.long_file_name // full original filename
    };
    std::fs::write(name, &attach.payload_bytes).unwrap();
}

Detecting and parsing embedded messages

Attachments with attach_method == 5 are nested .msg files (embedded messages). Use as_message() to parse them recursively:

let outlook = Outlook::from_path("email.msg").unwrap();

for attach in &outlook.attachments {
    if let Some(Ok(nested)) = attach.as_message() {
        println!("Embedded message subject: {}", nested.subject);
        println!("Embedded from: {}", nested.sender);
        // You can access all fields on the nested message, including
        // its own attachments (which may also be embedded .msg files)
    }
}

// Or use the convenience method:
for attach in &outlook.attachments {
    if attach.is_embedded_message() {
        println!("{} is an embedded .msg", attach.display_name);
    }
}

Inline images (Content-ID)

HTML bodies reference inline images via cid: URIs. Use content_id to resolve them:

let outlook = Outlook::from_path("email.msg").unwrap();
let mut html = outlook.html.clone();

for attach in &outlook.attachments {
    if !attach.content_id.is_empty() {
        // Replace cid: references with actual data
        let cid_ref = format!("cid:{}", attach.content_id);
        let data_uri = format!(
            "data:{};base64,{}",
            attach.mime_tag,
            base64_encode(&attach.payload_bytes), // your base64 encoder
        );
        html = html.replace(&cid_ref, &data_uri);
    }
}

RTF decompression and HTML extraction

Many .msg files store the body as compressed RTF rather than HTML. Use rtf_decompressed() to get the raw RTF, or html_from_rtf() to extract embedded HTML:

let outlook = Outlook::from_path("email.msg").unwrap();

// Get the best available HTML body
let html = if !outlook.html.is_empty() {
    outlook.html.clone()
} else {
    // Many messages embed HTML inside compressed RTF
    outlook.html_from_rtf().unwrap_or_default()
};

// Or work with the raw decompressed RTF directly
if let Some(rtf_bytes) = outlook.rtf_decompressed() {
    std::fs::write("body.rtf", &rtf_bytes).unwrap();
}

Named properties (MAPI 0x8000+ range)

The parser automatically resolves MAPI named properties — both well-known dispID-based properties (e.g. ReminderSet, InternetAccountName, AppointmentStartWhole) and custom string-named properties stored in the __nameid_version1.0 streams. These are merged into the same property maps used for standard MAPI properties, so they appear transparently in the parsed output and JSON serialization.

Message metadata

let outlook = Outlook::from_path("email.msg").unwrap();

// Timestamps (ISO 8601 UTC, empty string if unavailable)
println!("Delivered: {}", outlook.message_delivery_time);
println!("Submitted: {}", outlook.client_submit_time);
println!("Created:   {}", outlook.creation_time);
println!("Modified:  {}", outlook.last_modification_time);

// Classification
println!("Class:       {}", outlook.message_class);  // e.g. "IPM.Note"
println!("Importance:  {}", outlook.importance);      // 0=Low, 1=Normal, 2=High
println!("Sensitivity: {}", outlook.sensitivity);     // 0=Normal, 1=Personal, 2=Private, 3=Confidential

JSON output

let outlook = Outlook::from_path("email.msg").unwrap();
let json = outlook.to_json().unwrap();
println!("{}", json);

Available fields

Attachment fields

Methods

Performance

from_path and from_slice use an optimized zero-copy header parsing path (Reader::from_bytes) that avoids the double-allocation overhead of streaming through BufReader. For large .msg files this reduces peak memory usage and parse time compared to the generic from_reader path.

Requirements

• Rust edition 2024 (rustc 1.85+)

Running the example

cargo run --example parse-email
# or with a specific file:
cargo run --example parse-email -- path/to/email.msg

Running tests

cargo test

Contributions

Feel free to open pull requests to contribute, enhance, or fix bugs.

License: MIT