100% Offline Meeting Transcription In‑Room Speaker Identification Fully Customizable AI Summarization
Transcript Tab Visual
Placeholder: User scrolling around transcript and renaming speaker names (Add your screenshot or video here)
Powerful Speaker Detection and voice recording
Handles in-person meetings
Computational diarization using state-of-the art segmentation models tracks individual speaker turns locally.
Mic vs System channel isolation
Allows for further cleanup of overlapping audio, making local speech-to-text resolution extremely high quality.
No meeting bots
Records locally and privately on your machine without requiring external bots to join your calls.
Echo cancellation
Direct loopback captures filter speaker reflections, preventing audio bleed and echoes during calls.
Waveforms Visual
Placeholder: Recording waveforms during mic and system audio recording (Add your screenshot or visual here)
Scriptable AI Pipelines means Infinite Meeting Summary Possibilities For however you like to slice it.
Write lightweight scripts using the Rhai engine to parse transcripts, filter speakers, and map-reduce long conversations directly on your device.
fn run(ctx) {
let transcript = ctx.transcript;
let summary = ctx.summary;
let llm = ctx.llm;
summary.append("# 📋 Action Items & Deadlines\n\n");
llm.write(summary,
"Extract all action items, assignees, and deadlines " +
"from the following transcript. Output in a clear checklist.\n\n" +
"Transcript:\n" + transcript.text()
);
}fn run(ctx) {
let transcript = ctx.transcript;
let summary = ctx.summary;
let llm = ctx.llm;
summary.append("# 🎯 Executive Synthesis\n\n");
llm.write(summary,
"Summarize the key strategic goals discussed in " +
"this meeting. Focus on high-level decisions.\n\n" +
"Transcript:\n" + transcript.text()
);
}fn run(ctx) {
let transcript = ctx.transcript;
let summary = ctx.summary;
// Filter turns to isolate PM statements
let pm_turns = filter_turns(transcript.turns, "Product Manager");
summary.append("# 👤 Product Manager Directives\n\n");
for i in 0..pm_turns.len() {
summary.append("- " + pm_turns[i].text + "\n");
}
}
// Helper logic embedded in scripting engine
fn filter_turns(turns, speaker_name) {
let filtered = [];
for i in 0..turns.len() {
if turns[i].speaker == speaker_name {
filtered.push(turns[i]);
}
}
return filtered;
}📋 Action Items & Deadlines
- Sarah to deploy the staging environment by Tuesday morning.
- David to draft the user feedback review slides for the sprint demo.
- John to verify local Whisper model compatibility on Windows builds.
- Everyone to review the new brown color scheme overhaul and approve changes.
🎯 Executive Synthesis
Key Decisions:
- Approved the brand overhaul to the warm mocha/espresso theme.
- Positioned custom scripting as a flexible "Power Feature" to keep landing experience simple.
- Scheduled release pending code-signing confirmation.
👤 Product Manager Directives
- "We need to ensure local processing is completely airtight before shipping."
- "Our users care deeply about privacy — not having cloud bots join meetings is our major selling point."
- "Make sure the custom scripting engine feels intuitive but is marked as an advanced tool."
Analytics Dashboard Visual
Placeholder: Analytics dashboard showing speaker distribution and meeting metrics (Add your screenshot here)
Instant Meeting Analytics
Analytics are revealed instantly after transcription. Displays speaker talk-time, word counts, and meeting duration at a glance.
More Features
...and more coming
Quick Search Sessions
Locate specific terms, speakers, or topics across your entire meeting history instantly.
Multiple Export Options
Export transcripts, summaries, and lists in Markdown, JSON, or plain text to fit your workflow.
Familiar Chat Interface
A clean sidebar and session layout inspired by ChatGPT, Gemini, and Claude for natural navigation and interaction.
Modify Transcripts & Speakers
Edit text directly using text mode, and easily select and rename speakers or groups of speakers to clean up your logs.
Save Settings to Profiles
Save selected models, speaker counts, summarization scripts, and audio filter preferences to reusable profiles for different applications.
Samurai Scribe vs. cloud alternatives
See how a privacy-first approach stacks up against the competition.
| Samurai Scribe | Otter.ai | Fireflies | Whisper (CLI) | |
|---|---|---|---|---|
| Runs 100% locally | ✓ | ✗ | ✗ | ✓ |
| No audio uploaded | ✓ | ✗ | ✗ | ✓ |
| Works offline | ✓ | ✗ | ✗ | ✓ |
| Speaker detection | ✓ | ✓ | ✓ | ✗ |
| AI summaries | ✓ | ✓ | ✓ | ✗ |
| Scriptable pipelines (Rhai) | ✓ | ✗ | ✗ | ✗ |
| Desktop GUI | ✓ | Web only | Web only | CLI |
| One-time purchase | ✓ | $17/mo | $19/mo | Free (OSS) |
| No account needed | ✓ | ✗ | ✗ | ✓ |
One-time purchase.
Own it forever.
No subscriptions, no recurring fees. Pay once and get lifetime access including all future updates.
Free
Perfect for trying out Samurai Scribe using fast local Whisper Tiny & Base models.
- Unlimited transcription
- Unlimited AI summaries via built-in templates
- Mic vs system speaker grouping
- Echo cancellation & loopback isolation
- Record or upload audio
- 100% offline & local compute
- Access to all Whisper transcription models (Tiny to Large v3)
- Customizable AI summarization scripts
- Speaker diarization & identification (same room or channel)
- Lifetime Pro updates & license support
Pro
Full power with access to all Whisper models, speaker detection, and custom scripting.
- Unlimited transcription
- Unlimited AI summaries via built-in templates
- Mic vs system speaker grouping
- Echo cancellation & loopback isolation
- Record or upload audio
- 100% offline & local compute
- Access to all Whisper transcription models (Tiny to Large v3)
- Customizable AI summarization scripts
- Speaker diarization & identification (same room or channel)
- Lifetime Pro updates & license support
30-day money-back guarantee. No questions asked.
FAQ & System Specs
Everything you need to know about the offline transcription experience.
What does 100% offline mean?
It means everything runs locally on your own computer. All audio processing, transcription model inference (Whisper), speaker identification, and custom summaries are executed in memory on your local CPU or GPU. Absolutely no audio or text data is sent to the cloud. The only time the app communicates online is to verify a valid license key and to check for updates. Because our sustaining cloud server costs are zero, we pass those savings directly to you: buy once, use the app forever with no monthly subscription fees.
Are my conversations kept private?
Completely. Because Samurai Scribe runs 100% locally, your recordings, transcripts, summaries, analytics, and all other user data never leave your device. We have no cloud servers to store your meetings, no tracking scripts, and no access to your conversations. Your data is entirely yours.
What about summarization—is that local too?
Yes. Custom meeting summarization runs entirely locally. We currently support local LLM inference via Ollama, allowing you to connect your own local models (like Llama 3 or Mistral) to keep 100% control of your text data.
How do I learn to create a custom summarization script?
We are working to make our documentation better. In the meantime, we recommend reading the built-in, in-app templates to understand the basic structure. You can also try to paste the structure into your favorite AI service (such as ChatGPT, Gemini, or Claude) and ask it to help. It won't have all the answers about our API, but is great at making inferences.
How does it record if there are no meeting bots?
Instead of inviting external bots to join your virtual meetings, Samurai Scribe records your system audio and microphone feeds directly at the OS layer. This allows you to capture Zoom, Google Meet, Teams, or Slack calls privately without displaying a bot in the call. When you first launch the app, macOS or Windows may prompt you to grant the necessary audio recording and screen-capture (loopback) permissions to capture desktop audio streams.
What are the system requirements?
A hefty to moderate computer is recommended for the fastest, hardware-accelerated transcription results (such as Apple Silicon Macs or Windows PCs with discrete Nvidia GPUs). However, transcription is technically possible on almost any computer using CPU fallback. We invite you to download the free version, test it out on your machine, and decide for yourself.
macOS Details:
- Apple Silicon (M1, M2, M3, M4 family) recommended for hardware-accelerated transcription. Intel-based Macs are supported using CPU fallback.
- macOS 12 (Monterey) or higher.
- 8 GB RAM minimum.
Windows Details:
- Windows 10 or 11 (64-bit editions).
- 8 GB RAM minimum (16 GB recommended).
- Discrete Nvidia GPU (Nvidia CUDA supported) recommended for significantly faster local AI inference.
Storage Details:
- Approximately 1.5 GB of free storage to house the app executable and speech-to-text models.
How many active machines can I use my license on?
You can activate Samurai Scribe on up to 2 machines simultaneously. If you upgrade your computer or need to switch machines, you can deactivate the license from the app settings on your old machine to free up a slot for the new one.
Ready to take back your privacy?
Download Samurai Scribe for free. No account needed — just install and start transcribing.