Studio-grade desktop applications engineered for creative professionals, video editors, and audio engineers. Deterministic on-premises inference, dedicated GPU acceleration, and uncompromising confidentiality — free from cloud dependencies and subscription overhead.
100% on-premises execution, zero cloud telemetry, uncompromised confidentiality.
High-throughput tensor core pipeline engineered for discrete GPU architectures.
High-accuracy multilingual speech decoding and source-aware dialogue segmentation.
Predictable capital investment, permanent commercial ownership, zero recurring fees.
Purpose-built Windows applications engineered to drastically accelerate studio and creative workflows.
Studio-grade desktop speech-to-text and subtitling suite powered by OpenAI Whisper Large-v3, PyAnnote Audio, dGPU acceleration, and 100+ languages. v1830 BETA ready now; manual-documented first stable v1840 coming soon!
Lightweight, fully offline desktop utility: extracts every audio track (.wav) and subtitle track (.srt / .sup) from .mkv, .m2ts and .mts files in one click – no account, no internet connection.
Active engineering is underway for the first stable Whisper Transcriber Pro v1840 (detailed in the published user manual), alongside new dedicated studio utilities for Windows, macOS, and Linux!
Whisper Transcriber Pro is an industrial-grade workstation software for speech-to-text transcription and subtitling, merging the OpenAI Whisper Large-v3 neural transformer with PyAnnote Audio 3.1 acoustic clustering. Engineered for broadcast and studio workflows, it provides sample-accurate timecoding, Silero-VAD noise gating, and an integrated waveform editor — 100% on-premises with absolute data confidentiality.
The Pro licence key is digital content delivered immediately after payment. By purchasing, you expressly request immediate delivery and acknowledge that you thereby lose your statutory right of withdrawal: all purchases are final and non-refundable, and the price of a paid licence cannot be reclaimed in exchange for returning the licence key. Defects are corrected through updates; details: Terms and Conditions
Whisper Transcriber Pro processes nearly all world languages with native accuracy, including multi-lingual switching:
Real-time acoustic waveform timeline, interactive subtitle blocks, and sample-accurate timecode navigation.
Whisper Large-v3 model selector, Silero-VAD noise gating thresholds, and source language tuning.
Built-in compute diagnostic measuring discrete GPU tensor throughput and real-time VRAM allocation.
Dedicated graphics adapter selection, precision VRAM buffer sizing, and multi-core parallelism.
This purpose-built utility extracts every audio track (.wav) and subtitle track (.srt / .sup) from your .mkv, .m2ts and .mts files in one click – fully offline, no cloud, no account. A perfect companion for editors and subtitlers.
Founder Norbert Zsolt Szabó's journey from syntax barriers to natural language engineering.
Few people know this about me, but I majored in math and computer science in high school. Between 9th and 12th grade I was already deeply interested in the world of technology. Back then, Pascal was the curriculum, and I honestly admit that writing syntax was far from my strong suit, and I found the pace of development terribly slow, inefficient and boring. To me, the bandwidth of software development was a disappointingly slow and boring process, so at the time I was more interested in filmmaking and video editing. My strong ability to think in complex algorithms, however, proved to be a great strength even in my film projects. What captivated me even then was the internal logic of software, creating operational principles, and system-level design.
Over the years of directing and post-producing complex broadcast projects, I have utilised virtually every leading industry suite, frequently encountering bottlenecks rooted in architectural inconsistencies and ergonomic friction. In an exacting studio environment, deterministic reliability is not a luxury—it is fundamental: technology must transparently and reliably serve creative intent.
I have consistently maintained an acute diagnostic perspective on systemic anomalies. When examining any production pipeline, beyond cataloging surface errors, my focus immediately gravitates toward foundational structural optimisation and human-centred design. The objective is to engineer purpose-built tools that strictly conform to modern broadcast and high-throughput production standards.
The turning point came in 2024, when I discovered SI-supported development. It was a relief that syntactic difficulties are no longer an obstacle. I can finally focus on what really interests me: functions, logical structure, and user journeys.
I formulate the concept fluently in Hungarian or English, and the LLM translates it into Python. OpenAI's o1 model, and then the release of Gemini 3.1 brought the level where my instructions formulated with engineering precision result in immediately working code 99% of the time.
I perceive this modern paradigm as the symbiosis of high-level architectural governance and SI-assisted software engineering. Rather than expending intellectual energy on manual syntax mechanics, focus is directed toward neural transformer integration, high-throughput dGPU tensor pipelines, and broadcast-grade ergonomics. Collaborating closely with advanced frontier models allows an independent engineer to sustain production velocity, mathematical precision, and system robustness that historically required dedicated engineering departments.
Why does Windows display the blue warning on first launch, and how to start the app in 2 simple clicks?
Always download our applications (specifically Whisper Transcriber Pro and WR Media Extractor) strictly and exclusively from the official wraithdreams.com platform via the verified download links published here! Our software tools are currently NOT officially distributed on any third-party software repository, external mirror, or marketplace. Builds obtained from unauthorized external sites, file-sharing portals, or modified "cracked" repack releases present severe cybersecurity hazards and may carry viruses, trojans, or spyware. Wraith Dreams guarantees digital cleanliness and binary integrity solely for original installers downloaded directly from our official channels; we expressly assume zero liability for damages, data corruption, or system malfunctions resulting from unverified third-party packages or non-original installers.
Microsoft Defender SmartScreen prevented an unrecognized app from starting. Running this app might put your PC at risk.
This advisory banner is a completely natural false positive (false alarm)! Microsoft Defender SmartScreen displays this alert solely because our independent Wraith Dreams Software brand and newly released installer versions have not yet accumulated widespread telemetry signals across a massive volume of installed machines. Our tools are 100% clean, verified compiled binaries, completely virus-free, and safe — readily launched in two simple clicks below.
Every desktop application operates strictly on-premises as a compiled clean binary. It contains zero third-party trackers, no adware, and zero outbound network telemetry.
Microsoft Defender SmartScreen relies on cloud telemetry reputation metrics: newly released or updated binary builds trigger an advisory banner until accumulating a broad threshold of telemetry signals.
Official Microsoft Documentation: “Checking downloaded files against a list of files that are well known and downloaded frequently. If the file isn't on that list, Microsoft Defender SmartScreen shows a warning, advising caution.”This means the notification appears solely because the fresh build has not yet accumulated sufficient download frequency across machines, NOT because it contains malware.
Enterprise Extended Validation (EV) hardware certificates entail significant recurring overhead that we refuse to pass onto our users. Development resources remain strictly channeled into compute optimisation and direct engineer support.
As our applications are not officially available anywhere else, we strictly disclaim liability for any damage caused by unverified third-party copies, repacks, or modified installers.
Write directly to Norbert! Your feedback shapes the future versions of Whisper Transcriber Pro and upcoming tools.
