ForgeScribe icon
AI Audio

ForgeScribe

Local transcription app for Windows. Turn interviews, lectures and voice notes into text on your own PC - no uploads, no account, no subscription. $14.99 once, 7-day trial.

Use Tool
0
Uses
—
Verified
Open
Access
Nov 2026
Added
Back to Directory

Open ForgeScribe

This tool opens on an external website.

Use Tool

Found a Bug or Missing a Feature?

Report any problems or suggest new features you want. You can write in English, Urdu, Hindi, or any language! Our AI will analyze your feedback and upgrade the tool.

About This Tool

Local transcription app for Windows. Turn interviews, lectures and voice notes into text on your own PC - no uploads, no account, no subscription. $14.99 once, 7-day trial.


ForgeScribe: Transcription That Never Leaves Your PC

Most transcription tools start the same way: upload your recording to someone else's server and wait for the text to come back. That works fine for a podcast clip. It feels different when the recording is a job interview, a therapy session, a business meeting or a family story told once. ForgeScribe is built for that second feeling. It is a Windows app that turns recordings into text entirely on your own computer, with no uploads, no account and no subscription.

The app comes from BrutalFoundry and is sold through the Microsoft Store for a one-time $14.99, with a fully unlocked seven-day trial. You download the speech models you need once, and after that the recognition runs locally, even offline. Your recordings and your transcripts stay where they were made.

Checked on November 12, 2026. Prices and features below come from ForgeScribe's own product page on that date.

Why local transcription matters

Privacy is the obvious reason, and it is a real one. A recording sent to a cloud service passes through hands you cannot see, under retention rules you did not write. With ForgeScribe there is nothing to upload in the first place: the app processes your media on your PC, and ForgeScribe does not upload your media or transcripts at all. There is a practical side too. No account means no password to manage and no service to sign into. No subscription means the price you pay is the price, and the app keeps working the same way next year. And because the models live on your disk, a long recording costs you nothing per minute, which is not true of most cloud transcription, where the meter runs the whole time.

From recording to finished text in four steps

The workflow is plain. First you add a recording: choose audio or video files, add a batch of them, or record from a selected microphone or your system-audio output. Second, you choose a model: Fast, Balanced, Accurate or Best, each downloaded once. Compatible Vulkan hardware can speed things up, and CPU transcription is available when it cannot. Third, you review and edit the timed transcript: search and replace text, split or merge sections, and rename the optional speaker labels. The app is honest that recognition and speaker labels can be wrong, especially with noise or overlapping speech, so this step is part of the process, not an afterthought. Fourth, you export. While you record, live caption previews show the words as the audio arrives, though those are a provisional preview rather than the final transcript, and the original recording is always saved independently.

The models, and the hardware they need

Four built-in models cover the usual trade-off between speed and accuracy. Fast gets through a recording quickly, Best takes longer and aims higher, and the two in the middle split the difference. You download each model once and then use it locally as often as you like. If you already have a compatible model of your own, ForgeScribe can import a complete whisper.cpp GGML .bin model from a source you trust, and it validates a local copy before selecting it. On requirements, you need an x64 PC with Windows 10 version 22H2 or Windows 11, plus disk space for the models, your recordings and decoded audio, and a microphone if you want to record from one.

What you can do with the result

A transcript is only useful if it can leave the app, and ForgeScribe exports generously: TXT, SRT, VTT, DOCX, Markdown, CSV, JSON, LRC and PDF. Save text for your notes, DOCX for editing, or timed subtitles for a video workflow. SRT and VTT exports make it a practical subtitle tool as well: review the words and the timing, then drop the subtitle file straight into your editor. One honest note from the product page is that PDF has limited character coverage, so if the app reports unsupported text, DOCX or TXT is the better export. Twelve menu languages are supported, including English, Spanish, French, German, Hindi, Arabic and Urdu, though some technical messages and model descriptions remain in English.

What it costs, and what the trial covers

The configured US price is $14.99 once, with a fully unlocked seven-day trial through the Microsoft Store. Microsoft shows the price and availability for your region and manages purchases and licensing. One thing to know before you buy: an active Store license is required to open ForgeScribe, so when a trial expires you need to purchase to open the app and your saved transcript history. Your original recordings and any files you already exported stay usable outside the app either way. There is no subscription at any point.

The honest limits

Local transcription is not magic, and the product page says so plainly. One example is not an accuracy or speed guarantee: results depend on the recording, the language, the model and the hardware. Noisy audio and people talking over each other will produce errors, and speaker labels are a best effort. The app is Windows-only, so Mac and Linux users need a different tool. Transcription can be interrupted, though saved checkpoints can resume it, and recording WAV headers are updated as audio is written; even so, the page advises keeping the original media, because recovery is not guaranteed from every failure. If you need a quick one-off transcript in a browser instead, the web-based AccurateScribe online transcription tool and WhisperTranscribe AI transcription software both run without installing anything, and our own speech to text tool converts speech in the browser for short jobs. Going the other way, the text to speech tool turns written words back into spoken audio.

Local app or online transcriber?

The choice comes down to what you value most. Pick the local route when the recording is sensitive, when you transcribe often enough that per-minute pricing would add up, or when you want to work offline. Pick an online transcriber when you need one quick transcript on a borrowed computer, when you are on a Mac or a phone, or when installing software is more trouble than the job is worth. Many people end up with both: the local app for the recordings that matter, and a web tool for the quick ones. ForgeScribe's case is strongest exactly where privacy and volume meet, which is also where its one-time price makes the most sense.

Frequently asked questions

Does ForgeScribe upload my recordings?

No. Speech recognition runs locally on your PC after a one-time model download. ForgeScribe does not upload your media or transcripts. Microsoft manages purchases and licensing and may require an internet connection, but your recordings stay local.

What does ForgeScribe cost?

US$14.99 once, with a fully unlocked seven-day trial through the Microsoft Store. There is no subscription. Microsoft shows the price and availability for your region.

Which versions of Windows does it need?

An x64 PC with Windows 10 version 22H2 or Windows 11. You also need disk space for the models, recordings and decoded audio, and a microphone if you want to record from one.

Can ForgeScribe make subtitles?

Yes. Review the words and timing, then export SRT or VTT for a video workflow. TXT, DOCX, Markdown, CSV, JSON, LRC and PDF exports are available too.

What languages does it support?

Twelve menu languages, including English, Spanish, French, German, Hindi, Arabic, Bengali, Portuguese, Russian, Indonesian and Urdu. Spoken-language selection is separate, and speech translation produces English text.

What happens if transcription is interrupted?

Saved checkpoints can resume an interrupted transcription, and recording WAV headers are updated as audio is written. These recovery features do not guarantee recovery from every failure, so keep the original media.

More Tools in AI Audio