Most dictation tools hand you a transcript. Pithflow hands you finished text. Hold your hotkey, talk, let go — polished text lands in Slack, Gmail, VS Code, any Windows app. Cleanup strips the ums, fixes punctuation, applies your tone. • Windows-native, not an Electron port • Works over Citrix and RDP — nothing installs in the remote session • Bilingual EN/ES, Spanglish included, detected per dictation • Vocabulary packs for medical, legal, engineering terms Free tier, no card. Pro $9.99/mo.
Hey Product Hunt 👋
I built Pithflow because I talk faster than I type, but every dictation tool I tried handed me a wall of raw transcript that I then had to clean up by hand. That isn't faster — it just moves the work somewhere else.
So Pithflow makes the cleanup part of the pipeline. Hold a key, talk, let go, and what lands in the text field is already punctuated, de-ummed and in the tone you picked. Straight into whatever app you were already in: Slack, Gmail, VS Code, a CRM field.
It's Windows-only, on purpose. Dictation on Windows has been the neglected case for years — the good tools went Mac-first and showed up here late as Electron ports. Pithflow is a native Windows app, so it isn't hauling a browser engine around just to listen to you.
Two things that came straight out of my own workflow:
Remote desktops. Nothing installs inside the Citrix/RDP session — the audio is captured on your local machine — so it works in locked-down environments where you can't install anything on the far end. That turned out to matter to far more people than I expected.
Spanish and English in the same sentence. I work bilingually all day, and most tools make you pick a language in settings and then mangle the code-switching anyway. Pithflow detects per dictation.
There's also a personal dictionary and vocabulary packs for the fields where generic speech-to-text falls apart — it is never going to spell "metoprolol" on its own.
Free tier is 2,000 words a week, no card. Pro is $9.99/mo past that.
I'd love to hear what breaks it. Specifically: what words does your current dictation tool always get wrong? That's the list I want to fix next.