Olib AI
In DevelopmentVideo
macOS Studio

CamScribe

Native macOS screen and camera studio with on-device intelligence

CamScribe is a native macOS recording studio and non-destructive editor built for Apple Silicon. It records screen, window, region, camera, microphone, and application-isolated audio onto independent tracks with crash-recoverable chunks. The studio integrates a syllable-paced click-through teleprompter, live face detection, and post-recording median face centring. All video processing, rendering, and speech transcription run on your Mac using Core ML and Metal. CamScribe also includes a local loopback MCP server with 45 tools, allowing AI agents to edit and assemble video timelines directly.

Visit CamScribeBook Technical Walkthrough
Deployment SpecHardware First
Platform RequirementsmacOS 15 or later, Apple Silicon (M1/M2/M3/M4)
Graphics & ComputeMetal-accelerated pipeline shared by preview and export
Capture SourcesDisplay, application window, custom region, camera, mic, isolated system audio
Export FormatsMP4 (H.264 / HEVC) up to 4K resolution
Zero Telemetry Standard:No marketing trackers, no remote analytics telemetry, and complete data ownership.
4K
Master Export
MP4 H.264/HEVC exports across 16:9, 9:16, 1:1, and 4:5 aspect ratios.
45 Tools
Local MCP Server
Exposes full editing, cutting, and captioning controls to AI agents.
4,344
Integrated Glyphs
Searchable icon library and 51 title card templates.
0 KB
Cloud Upload
All transcripts, timeline journals, and assets stay strictly on disk.
Specifications verified from source build
Recording Engine01

Multi-Track Crash Recovery

Separate audio and video streams write to segmented, atomic disk chunks. Even during unexpected system shutdowns, recorded tracks remain fully recoverable.

Presenter Tools02

Syllable-Paced Teleprompter

Calculates natural reading pacing based on syllable counts, keeping talking points at camera eye-level while remaining invisible in final captures.

Timeline03

Sparse Automation Timeline

Non-destructive editing with an absolute integer timebase. Unified automation primitives handle pan, zoom, blur, spotlight, and audio ducking.

Local AI04

On-Device Speech Transcription

Offline Core ML models generate accurate transcripts and render timed captions directly onto picture bands with zero network calls.

Core Engineering Capabilities

Detailed feature checklist for CamScribe, designed and audited for stability in real operational environments.

Multi-track capture of display, camera, microphone, and application audio

Crash-safe chunked recording prevents file corruption during sudden power loss

Syllable-paced teleprompter keeps eye contact aligned with the lens

Non-destructive multi-track timeline with unified visual automation primitives

Local Core ML transcription with timed caption rendering

Local MCP server enabling automated AI video assembly and timeline operations

Part of the Olib AI Ecosystem

Every product we ship serves as concrete proof of the systems and architectures we build for clients.

Flagship Platform

steffi.ai

Steffi reads inboxes, updates systems, processes documents, works in browser portals, and sends exceptions to a person. Your team controls access, spending, and approvals.

View specifications →
Anti-Bot Engine

Owl Browser

Tell it what to do in plain English, like 'click the login button,' and run hundreds of sessions at once. A drop-in replacement for Playwright and Puppeteer with 182+ built-in tools.

View specifications →
iOS Privacy

StealthOS

19+ apps, encrypted storage, tracking and phishing protection, and an on-device assistant designed to keep core work on the phone.

View specifications →
Custom Engineering

Need CamScribe customized for your enterprise?

Every capability in our products can be tailored and deployed inside your private infrastructure. You own the code, the models, and the data.

Book a founder consultationBrowse all products