CLI — src/index.ts

bun src/index.ts to-md <file> [--format <fmt>] [--at <iso>] [--out <file>]
bun src/index.ts from-md <file.md> --target <fmt> [--out <file>]
bun src/index.ts convert <file> --target <fmt> [--format <fmt>] [--out <file>]
bun src/index.ts to-html <file.md> [--out <file>]
bun src/index.ts --help | -h
bun src/index.ts --version | -v

node src/index.ts … takes the same arguments. So does deno run --allow-read --allow-write src/index.ts …. See Runtimes.

Streams

• With no --out, the output goes to stdout, so the tool works in a pipeline. With --out, it goes to that file only. The file's bytes match what stdout would have carried.
• Diagnostics and the loss count go to stderr.
• A lossy conversion still exits 0. The count is printed on stderr as convert: N loss recorded or convert: N losses recorded.
• No ANSI escape is ever written to stdout.

Exit codes

Flags

to-md

Document to Markdown. Wraps toMd. Input formats: docx, odt, xlsx, ods, pptx, odp, pdf. The format comes from the file extension unless --format is given.

$ bun src/index.ts to-md tests/fixtures/sample.docx --at 2026-01-01T00:00:00.000Z --out sample.md
$ echo $?
0
$ tail -3 sample.md
### Chunking

Sections are stable anchors for the retriever.

Without --out, the same 674 bytes go to stdout. The front matter and body are shown in toMd.

$ bun src/index.ts to-md note.md
convert: oconv: unsupported format (supported: docx, odt, xlsx, ods, pptx, odp, pdf)
$ echo $?
2

--at is checked before the file is even opened — a bad value or a missing input file both exit 2, but the --at message wins:

$ bun src/index.ts to-md tests/fixtures/sample.docx --at not-a-date
convert: --at: invalid timestamp "not-a-date": expected an RFC 3339 date-time such as 2026-01-01T00:00:00.000Z
$ echo $?
2

from-md

Markdown to a document. Wraps fromMd. Targets: docx, odt, pdf. --target is required.

note.md below holds # Hello, A *short* note. and a - [x] shipped task item.

$ bun src/index.ts from-md note.md --target docx --out note.docx
convert: 1 loss recorded
$ echo $?
0
$ bun src/index.ts from-md note.md --target pdf | head -c 8
convert: 1 loss recorded
%PDF-2.0

The loss is the task checkbox (list/task-marker-dropped), which neither target can carry.

$ bun src/index.ts from-md note.md
convert: from-md requires --target
$ echo $?
2
$ bun src/index.ts from-md note.md --target xlsx
convert: oconv: unsupported target (supported: docx, odt, pdf)
$ echo $?
2

convert

Document to document, with no Markdown step in the output. Wraps convert. Pairs: docx>odt, odt>docx, docx>pdf, odt>pdf. --target is required.

$ bun src/index.ts convert tests/fixtures/sample.docx --target odt --out sample.odt
$ echo $?
0
$ bun src/index.ts convert tests/fixtures/sample.docx --target docx
convert: oconv: unsupported pair (supported: docx>odt, odt>docx, docx>pdf, odt>pdf)
$ echo $?
2

to-html

Markdown to an HTML fragment. Wraps toHtml: safe and sanitised (raw HTML dropped, dangerous URL schemes neutralised, fw sanitiser pass). There is no loss count: the render drops raw HTML and unsafe URLs (see toHtml) but does not itemise them.

$ bun src/index.ts to-html note.md
<h1>Hello</h1>
<p>A <em>short</em> note.</p>
<ul>
<li> shipped</li>
</ul>

Other errors

$ bun src/index.ts
convert: missing command (to-md | from-md | convert | to-html)
$ echo $?
2
$ bun src/index.ts bogus x
convert: unknown command: bogus
$ echo $?
1
$ bun src/index.ts to-md
convert: to-md requires a file argument
$ echo $?
2
$ bun src/index.ts to-md missing.docx
convert: ENOENT: no such file or directory, open 'missing.docx'
$ echo $?
1