Record your screen, narrate your reasoning, and Mnemori turns it into documentation, coaching reviews, and clean notes. The thinking you voice while building becomes the artifact your team can learn from.
Start a recording, talk through what you're doing and why, then stop. Mnemori handles the rest.
Start a screen recording and narrate your work. Talk through decisions, explain why you chose one approach over another, flag things that feel tricky. Mnemori captures both the screen and your voice.
Mnemori sends your audio to Whisper and gets back a clean transcript — timestamped, segmented, and ready to work with. Your raw narration becomes searchable text.
Choose what you need. An SOP someone else can follow. A methodology doc explaining your design decisions. A coaching review surfacing where you're growing. Clean notes that read like you sat down and wrote them.
Each recording can produce any combination of these artifacts. Same transcript, different lenses.
A step-by-step procedure someone else can follow without the video. Prerequisites, exact steps, warnings about what's easy to get wrong.
The architectural rationale — what you built, why you built it that way, what trade-offs you weighed, and how the pieces connect.
A skill coach's analysis of your session. Strong moves, growth edges, concepts to reinforce, and concrete next steps for practice.
Your narration stripped of filler words and false starts, structured into readable paragraphs. Your thinking, just cleaner.
Other tools watch what you do and infer what happened. Mnemori listens to what you say and knows what you were thinking.
Clicked "Settings." Selected "Users." Pressed "Add."
Records actions. Infers intent. Produces generic step-by-step guides that document the what but lose the why.
"I'm adding them as a viewer first because they shouldn't see billing until we've scoped the project."
Captures reasoning. Preserves intent. Produces documents that explain decisions, not just procedures.
Mnemori is local-first. Recordings, transcripts, and artifacts live on your machine. Nothing touches our servers. The only external communication happens when you explicitly transcribe or generate — and that goes directly to the API provider, not through us.
API keys encrypted at rest via your OS credential store (DPAPI / Keychain)
Files overwritten with zeros before unlinking. Configurable auto-retention.
Every security-relevant action is logged locally. Credential values never recorded.
Strict Electron sandboxing. The UI cannot access your filesystem or OS directly.
No analytics, no usage tracking, no background sync. We don't know what you record.
Building toward certification. Compliance documentation maintained alongside every release.
Free during early access. Bring your own OpenAI and Anthropic API keys. Windows and macOS.
Create an account to get your download links.
You're in. Choose your platform.
You can also download again anytime by signing in.
Mnemori uses OpenAI for transcription and Anthropic for document generation. You provide your own API keys from those services, which means your data goes directly from your machine to theirs — never through us. You can create keys at platform.openai.com and console.anthropic.com.
Mnemori itself is free during early access. You pay only for the API calls — typically a few cents per recording for transcription and a few cents more for each generated document. A heavy day of recording might cost $1–2 in API usage.
Your recordings, transcripts, and generated documents live on your computer — not on our servers, not in the cloud. Mnemori works offline after you've signed in once. The only time data leaves your machine is when you explicitly press Transcribe or Generate, and that goes directly to the API provider.
Eventually, yes. We're exploring team features — shared projects, organization-wide coaching insights, managed API keys — that would sit behind a subscription. The core recording and generation workflow will remain available.
Windows 10+ and macOS 12+. Both are available during early access. Linux is on the roadmap.