Convert Outlook to Text Online Free

Outlook MSG to Plain Text: Read It, Index It, Search It

Someone forwards you a file called RE_ Invoice query.msg. You double-click it and your machine offers to open it in a text editor, which produces a screenful of binary garbage with a few legible words scattered through it. You don't have Outlook. You just want to know what the email says.

That's the shortest reason to convert Outlook MSG to text online: it's the fastest way to open a .msg file without Outlook and actually read the thing. The longer reason is that flat text is the format every search index, embedding model, and classifier wants — so if you're working through a mailbox export rather than a single message, this is where you start. Free, no sign-up, 50 MB per file, and nothing is stored on our side.

Why the File Won't Open

.msg is Outlook's own single-message format, and it isn't text. It's a compound binary file — a container with an internal directory tree, where the subject, the body, each recipient, and each attachment sit in separate named streams tagged with MAPI property identifiers. That structure is why a text editor shows you noise: you're looking at the raw bytes of a small filesystem, not at an email.

The open alternative is .eml, defined by RFC 5322, which really is plain text — headers, blank line, MIME body — and which any editor will show you. Thunderbird, Apple Mail, and most webmail exports hand you those. Outlook hands you .msg, and it's what you get when someone drags a message from their inbox onto the desktop or uses File > Save As with "Outlook Message Format" selected. Nothing is wrong with the file. You're just missing the one program that reads it natively, and converting it sidesteps that entirely.

What Flat Text Is Actually Better At

This isn't "Markdown with the formatting removed". Text is the correct output for a specific and fairly large set of jobs, and in most of them structural markup is a liability rather than a loss:

• Search indexing. Elasticsearch, SQLite FTS, Postgres tsvector — every one of them wants raw text in a column. Feed them Markdown and you've added a sanitising pass, plus a stemmer that now has to decide what to do with stray pipes and asterisks.
• Keyword sweeps and e-discovery. Convert a folder of correspondence, then grep it for a supplier name, a defined contractual term, or a phrase that shouldn't be in there. Offline, instant, and reviewable by anyone with a shell.
• Embeddings. Chunk, embed, retrieve by similarity. Formatting characters carry no meaning but still occupy positions in the vector — they dilute. Clean prose chunks retrieve better.
• Classification and sentiment. Routing support mail by topic, scoring customer replies for frustration, tagging a quarter's correspondence by theme. Bag-of-words and TF-IDF approaches will cheerfully treat ### as a token unless you strip it, so don't create it.
• Token economy at volume. Processing one email, markup costs nothing worth mentioning. Processing four thousand, every redundant character is multiplied by four thousand. Text is the floor.

When you're reading one thread rather than processing many — reconstructing who agreed to what, and when — the labelled headers and visible quoting levels of MSG to Markdown are worth the extra characters. The format toggle at the top of this page swaps between the two and keeps whatever file you've already selected.

Working Through a Mailbox Export

Nobody converts one .msg for a search corpus. The usual shape of this job is a few hundred messages dragged out of Outlook into a folder, which is where the batching matters: bundle them into a single archive and run ZIP to text, and every supported file inside converts in one upload rather than several hundred.

Email is often personal or commercially sensitive, so check your organisation's rules on handling correspondence before uploading work mail. Files are converted and returned, not retained.

The other thing to plan for is duplication, and it's the single biggest source of bloat in an email corpus. Replies are cumulative: message twelve contains messages one through eleven quoted inside it. Convert a whole thread and the opening message appears in your output a dozen times. Layer on the signature block — name, title, phone, office address — repeated on every message, plus the automatic legal disclaimer many companies staple to every outbound reply, and the boilerplate can outweigh the substance.

For search and embeddings that repetition is actively harmful, because the same passage gets indexed a dozen times and skews your retrieval scores. Two fixes worth the effort: keep only the newest message per thread when you want history, and write one regex to cut everything from the first "From:" quoting header or disclaimer opening line onward. The token counter under the output tells you how much you saved.

Artefacts You'll See in the Output

• HTML bodies flatten hard. Most business email is sent as HTML. Bold, bullet lists, coloured status tables in project updates — all of it becomes unstyled prose. Usually fine. If a message's meaning lived in a table, the numbers survive but the columns they belonged to don't.
• Quoting markers are inconsistent. Outlook writes a "From/Sent/To/Subject" block above the quoted text. Gmail writes "On [date], [name] wrote:". Older clients use > prefixes. A thread that crossed several organisations will contain all three conventions, which matters if you're writing a regex to split messages apart.
• Timestamps read as sent, not as your timezone. Sort a converted set chronologically and check the offsets before you trust the ordering across regions.
• Attachments aren't extracted. The .msg holds them, but the conversion gives you the message content. Save the attachment out separately and run it through PDF to text or Word to text if you need what's inside it.
• Pasted screenshots don't come through. Images in a message body aren't read, so an email whose entire content was a picture of an error dialogue converts to nearly nothing. Save the picture out and run it through image to text, which does recognise text in an image file.

Jobs This Solves

• Reading a message you were sent. The plainest case, and the most common. You need the words, you don't have a mail client that opens the format, and you're done in ten seconds.
• Building a searchable archive. Years of project correspondence into a folder of text files that grep, ripgrep, or a desktop search index can reach — no PST to mount, no mail client involved.
• Review and disclosure work. Sweep a set for defined terms, produce a hit list, and hand reviewers something they can read without licensing Outlook for everyone on the team.
• RAG over correspondence. Embed a customer's email history so a support assistant knows what was already promised, rather than contradicting a colleague from four months ago.
• Classifying at volume. Run sentiment or topic models across a quarter of inbound mail to find where escalations cluster.
• Feeding an assistant the background. When you need a model to know what was agreed before it drafts a reply, the thread as text is the cheapest way to give it that.

Frequently Asked Questions

How do I open a .msg file without Outlook?

Upload it here. The MSG container is parsed and the message comes back as readable plain text — sender, recipients, subject, date and body — with no Outlook, no plugin and no install. Free, no sign-up, 50 MB per file, nothing stored on our side.

Why does a .msg file look like binary garbage in a text editor?

Because it is a compound binary file, not a text file. Outlook stores a message as a container with an internal directory tree, where the subject, the body, each recipient and each attachment sit in separate named streams tagged with MAPI property identifiers. Opening it as text shows you the container's raw bytes with a few legible words scattered through.

Does it extract attachments?

No — the output is the message itself: headers and body. Attachments stay inside the file. If you need the contents of an attached document too, save it out of the message first and convert it on the page for its own format, or bundle several messages into a ZIP and use ZIP to text.

Can I convert a whole mailbox export?

Not in one upload — this takes a single .msg. For a folder of them, zip the folder and convert the archive, which walks every message in one pass. That is the usual route for discovery work, thread reconstruction, or building a searchable index over a mailbox export.

Is the email content confidential?

The file is processed and discarded — not stored, logged or indexed, and no account is attached to it. That said, if the messages are genuinely sensitive, the conservative option is the local folder converter, which reads files in the browser without an upload step.

The Rest of the Toolkit

MSG is one of the formats File2Txt handles, so you can send anything at it from a single upload instead of picking a page. Neighbours worth knowing: HTML to text for saved web pages and newsletter markup, and CSV to text for the exports that tend to arrive attached to the emails you're already converting. For live pages, Web2Txt pulls a URL straight into text, and on the code side there's a GitHub repository to text converter for pairing a thread with the implementation it was arguing about. The guide to preparing documents for LLMs goes through the general reasoning.

Repo2Txt is built and maintained by v12hero, an independent developer building privacy-first native and web apps.