Feature Request: `--stdin` support

[knwoop]

What

Support reading Markdown from stdin.

some-command | mo --stdin
some-command | mo --stdin --name "result.md" --target review

Why

I often pipe Markdown output (e.g. plans, diffs formatted as Markdown) from CLI tools to preview in a browser. Currently I write to a temp file first:

some-command > /tmp/out.md && mo /tmp/out.md

Since mo already handles in-memory content via drag-and-drop (without live-reload), --stdin could work the same way — read until EOF, send to the running server, done.

Proposed behavior

• Read stdin until EOF, add content to the session
• --name to set the sidebar display name (default: stdin-<timestamp>)
• Same name replaces existing content
• No live-reload (same as drag-and-drop)
• Non-blocking, consistent with existing background-server behavior
• Works with --target for grouping

[knwoop]

If this fits mo's direction, I'd be happy to work on a PR!

[k1LoW]

Thanks for the proposal! Nice feature!!

STDIN detection

Instead of requiring a --stdin flag, mo could detect stdin automatically when no file arguments are given and stdin is a pipe (not a terminal). This follows the common Unix convention:

some-command | mo
some-command | mo --name "result.md" --target review

Default name

stdin-<first 7 hex chars of content hash> instead of stdin-<timestamp>. Uploaded files are deduplicated by content hash within a group, so a timestamp-based name would cause a mismatch: piping the same content again would return the existing entry with the old timestamp name, not the new one. A content-hash name keeps the name and dedup behavior consistent.

--name and "same name replaces content"

This is tricky with the current design. File IDs are derived from content hash, so updating content would change the ID and break deep links and frontend references.

What do you think?

[knwoop]

Thanks for the feedback! All three points make sense!

• I was actually thinking stdin auto-detection via pipe would be better than an explicit flag too. Much cleaner.
• Content-hash-based default name makes sense to keep it consistent with dedup.
• Happy to drop --name and content replacement given the current ID design.