Your recordings stay yours
Audio and video are processed on your Windows computer instead of being uploaded to a transcription service.
Private AI transcription for Windows
Your conversations don’t belong in the cloud. With Vidrick, they never go there.
Transcribe audio and video directly on your Windows PC. No cloud uploads. No external AI servers. Just private, local transcription with automatic speaker separation and clean, readable results.
✓ Windows desktop app
✓ 7 days free · No credit card
✓ No automatic charge
A quick walkthrough
From recording to a transcript you can use.
Create an account, sign in to activate your trial, then download Vidrick on your Windows PC. Your audio is processed locally; internet access is needed for account setup, downloads, and license checks.
1. Choose your recording
Open an audio or video file, select the language and speaker options, and start transcription. Processing time depends on your recording, settings, and hardware.
2. Review the speaker-labeled transcript
Read through the result in the editor. Check speaker assignments, names, and overlapping speech against your recording, and correct anything that needs attention.
3. Export and put it to work
Export your reviewed transcript as TXT or Word, or use SRT/VTT for subtitles. Start with a short recording to find the settings that work for you.
Welcome to Vidrick
A clear introduction to the private transcription workflow.
When Vidrick opens for the first time, the welcome screen highlights the core experience: private local processing, speaker identification, conversation reconstruction, and professional export options.
Transcribe
Drop in your recording. Get a private transcript.
Choose your file, language, output format, and speaker count, then let Vidrick process it locally. Professional users can use compatible NVIDIA GPUs for dramatically faster transcription.
- Audio and video support
- 57 selectable languages plus Auto-detect
- Translate supported speech to English
- Automatic offline speaker diarization
- CPU processing on Personal · GPU/CUDA on Professional
- Progress tracking and estimated completion time
Clean output
Clean transcripts built for real conversations.
Vidrick turns raw speech into organized, readable conversations and saves the formats you choose. Your recordings and transcript workflow stay on your computer.
- Clean, speaker-organized output
- Local processing from start to finish
- One-click access to output files
- TXT, DOCX, SRT and VTT exports
Transcript Editor
Your transcript stays useful after transcription is finished.
Review completed transcripts, make corrections, and save the finished version without sending your conversation to another service.
AI Insights
Turn long conversations into useful takeaways.
Pull the important information out of a transcript with structured summaries, key points, action items, owners, and deadlines — inside the same private desktop workflow.
- Structured conversation summaries
- Action items and key points
- Private desktop workflow
- No cloud upload of your recordings
History
Your transcription history stays close to the work.
Quickly find recent jobs, see how they were processed, and jump back to source files and output folders from your local history.
- Recent job history
- Processor details
- Source and output paths
- Local record keeping
About & Updates
Privacy isn’t a setting. It’s how Vidrick works.
See your installed version, check for updates, review privacy information, and learn what powers Vidrick — all from inside the app.
Know who said what
Automatic offline speaker diarization separates conversations into clear, organized speakers.
Faster with your NVIDIA GPU
Professional unlocks GPU/CUDA processing on compatible NVIDIA hardware for much faster transcription.
Use the transcript anywhere
Export TXT and DOCX transcripts or create SRT and VTT subtitle files from the same workflow.
From raw speech to readable conversation
Because a transcript should read like a conversation, not a data dump.
Speaker 1: Yeah... Well... I think— Speaker 1: we should probably... Speaker 2: Okay.
→
Speaker 1:
Yeah, I think we should probably move forward.
Speaker 2:
Okay.
What’s better in 4.2
Faster processing. More consistent speakers. Cleaner conversations.
Vidrick 4.2 brings together faster CPU processing and stronger speaker intelligence for interviews, meetings, podcasts, and everyday conversations. Recurring voices stay more consistent, conversational handoffs are cleaner, and the entire workflow remains private on your Windows PC.
More efficient CPU processing
Hardware-aware CPU use and fewer unnecessary speaker-analysis passes help reduce processing time while preserving transcription and speaker accuracy.
Stronger speaker identity
Recurring voices stay more consistent throughout a conversation, helping reduce fragmented or duplicate speaker labels.
Cleaner speaker handoffs
Improved boundary analysis keeps brief replies, sentence endings, and rapid back-and-forth exchanges with the right speaker.
Built for real conversations
Interviews, meetings, podcasts, and multi-speaker recordings benefit from the same continuity-focused workflow, from short exchanges to long sessions.
Conservative duplicate handling
Potential duplicate identities are compared against sustained voice patterns and consolidated only when repeated acoustic evidence supports the match.
Automatic speaker count
Vidrick can infer the number of speakers without forcing a fixed count, while still preserving genuinely distinct participants.
Still completely local
All transcription, speaker identification, conversation reconstruction, insights, and exports continue to run on your Windows PC — your recordings stay yours.
Still private by design. Transcription and speaker processing continue to run locally on your computer.
Multilingual transcription
57 selectable languages plus Auto-detect.
Vidrick AI Transcriber 4.2 supports 57 named transcription languages plus Auto-detect. Choose the recording language directly, or let Vidrick detect it automatically.
- Includes English, Spanish, French, German, Portuguese, Arabic, Chinese, Hindi, Japanese, Korean, Russian, Ukrainian, Vietnamese and more
- Translate to English remains available
- Language selection stays inside the same local transcription workflow
Pricing
Private transcription without paying by the minute.
Start with 7 days of Professional. Keep the speed of GPU/CUDA or move to Personal for private CPU transcription.
PERSONALIndividuals
$6.99/month
Private, unlimited local transcription for individuals who don’t need GPU acceleration.
- Unlimited local transcription · no per-minute fees
- Automatic speaker diarization
- Clean conversation reconstruction
- TXT, DOCX, SRT & VTT exports
- Automatic updates
- Email support
MOST POPULAR
PROFESSIONALFaster local processing
$13.99/month
For professionals who want everything to stay local without giving up the speed of GPU-accelerated transcription.
- Everything in Personal
- Commercial use
- Priority support
- GPU/CUDA acceleration on compatible NVIDIA hardware
- Private processing with no cloud uploads
- Up to 3 activated computers
- 7 day free trial
BUSINESSTeams
Custom
Private AI transcription for teams and organizations that need licensing, deployment support, and a path to broader rollout.
- Multi-user licensing
- Team deployment
- Dedicated support
- Volume pricing
- Enterprise onboarding
Already using Vidrick?
Your subscription and devices are in one place.
Open the Customer Portal to manage your plan, billing, activated computers, account security, and the latest Vidrick download.