AI Scribe for Deep Work

Your work,
remembered
and understood.

Escribano watches your screen recordings and turns hours of deep work into structured, shareable summaries — so you never start a status update from a blank page again.

Free · Apple Silicon (M1+) · 16 GB RAM minimum

Independently built · No VC funding · GDPR compliant · Data stays in Europe
Wednesday session
25 February 2026 · 3h 12m
Auto-generated
Research
Reviewed Q4 competitor landscape and pricing benchmarks across three markets.
52 min
Client meeting
Discovery call with Müller & Partners — agreed on pilot scope for March.
45 min
Drafting
Drafted executive summary and revised intro section of the proposal doc.
38 min
Review
Annotated team's slide deck and left async comments for tomorrow.
22 min
"A focused afternoon — proposal work progressed significantly, and the Müller call opened a concrete next step."

Three steps from recording to clarity.

01

Download and install

Download the DMG, drag Escribano to Applications. On first launch, grant Screen Recording permission. The recorder starts automatically — no terminal, no config, no daemon to manage.

02

It watches your screen

Always-on capture at one-second intervals. A local VLM running entirely on your machine understands your activity — coding, debugging, reviewing, researching — and groups it into TopicBlocks automatically.

03

Ask for your artifact

Open the menu bar app, pick your format — standup, session card, or narrative — and Escribano generates it from your recent work. Copy and paste. Nothing leaves your machine.

Also open source.

The TypeScript pipeline powering Escribano's VLM analysis is fully open source. Use it directly for batch processing, or contribute to the engine.

Terminal
$ npx escribano --file session.mov
✓ Checking prerequisites...
✓ Loading VLM model (Qwen3-VL-2B-4bit)
✓ Extracting frames (adaptive sampling)
✓ Running VLM inference on 142 frames
✓ Transcribing audio with whisper.cpp
✓ Segmenting by activity
✓ Generating summary artifact
Done! Artifact saved to ~/.escribano/artifacts/

Batch pipeline

npx escribano --file recording.mov

Requirements

  • macOS Apple Silicon (M1/M2/M3/M4)
  • 16GB RAM minimum (32GB recommended)
  • ~10GB disk for models

What you get

  • 3 artifact formats: card, standup, narrative
  • Activity detection: coding, debugging, research, meetings
  • Audio transcription with whisper.cpp
  • All processing runs locally

For people whose work
is hard to explain.

Team Lead · Manager

The Quiet Orchestrator

You spend your day unblocking people, reviewing work, and making calls that don't produce visible output. Escribano gives your invisible work a voice.

Analyst · Consultant · Researcher

The Deep Thinker

Three hours of research, and you have six browser tabs, two insights, and a blank status update. Escribano remembers the journey, not just the destination.

Account Manager · Sales

The Relationship Builder

Calls, proposals, follow-ups — your day is relational and fast-moving. Let Escribano handle the documentation so you can stay present with people.

Developer · Engineer

The Builder

You debug, refactor, review PRs, and lose hours in terminal windows. Escribano understands the difference between coding and debugging — so your standup actually reflects what you shipped.

"Deep work deserves a record. Not surveillance — understanding."
— The Escribano Manifesto
🇪🇺

Built in Europe

Your data stays in Europe. GDPR isn't a checkbox — it's our default.

⚖️

No investors

Bootstrapped and profitable-first. We answer to our users, not a board.

🔒

Private by design

Local processing where possible. No training on your sessions. Ever.

🌿

Calm technology

No growth loops, no engagement tricks. Just a tool that does its job quietly.

Simple, honest pricing.

Local-first. No cloud required. Beta testers get unlimited history and direct access to the roadmap.

Free
Free
Always · local processing
  • 7-day history window
  • 3 artifact formats — card, standup, narrative
  • All processing runs locally
  • macOS Apple Silicon · 16 GB+ RAM
Download
All data processed in accordance with GDPR. European infrastructure only. We do not train on your sessions.