Cap is an open source alternative to Loom that enables users to record, edit, and share screen videos with AI-powered features like automatic transcription, summaries, and smart chapters. Designed for teams and creators who value privacy and control, it offers native macOS and Windows apps alongside a web interface, with full self-hosting capabilities using Docker, Railway, or Coolify.
Built as a monorepo with Tauri (Rust) for desktop, Next.js for web, Drizzle ORM for MySQL, and Tinybird for analytics, Cap supports custom S3 storage, background uploads, and local recording. Its architecture ensures data sovereignty while maintaining a polished, fast user experience across platforms.
What You Get
- AI-Powered Summaries - Automatically generates concise video summaries and titles using AI after recording.
- Auto Transcription - Creates accurate text transcripts of all recordings with timestamped chapters.
- Smart Chapters - AI segments recordings into labeled sections (e.g., ‘Introduction’, ‘Project Setup’) for easy navigation.
- Custom S3 Integration - Connect your own S3 bucket to store recordings, ensuring full data control and compliance.
- Studio Mode - Records 4K/60fps video locally with customizable backgrounds, padding, rounded corners, shadows, and cursor effects.
- Screenshot Mode - Capture screenshots with one hotkey (⌘+⇧+S), then enhance with annotations, gradients, padding, and shadows before sharing.
- Cross-Platform Native Apps - High-performance desktop apps built with Tauri (Rust) for macOS and Windows — no Electron bloat.
- Self-Hosting Support - Deploy Cap entirely on your infrastructure using Docker Compose, Railway, or Coolify with full config options.
- Background Uploads - Videos upload while you record, enabling instant sharing without waiting for processing to finish.
- Comment & Reaction System - Viewers can leave text comments and emoji reactions (😂, 👍, 👎) directly on shared videos.
Common Use Cases
- Running a developer documentation channel - A tech writer uses Cap to record step-by-step tutorials, auto-generates transcripts for SEO, and hosts videos on their S3 bucket to avoid third-party platform lock-in.
- Internal team onboarding - A remote engineering team records Loom-style walkthroughs of internal tools, self-hosts Cap on a VPS, and shares password-protected links with new hires.
- Product demo creation - A SaaS founder uses Studio Mode to record polished 4K product demos with custom backgrounds and cursor effects, then shares them directly with investors via Cap.link.
- Compliance-sensitive environments - A healthcare or finance team deploys Cap on-premises to record training videos without sending data to external servers, meeting GDPR and HIPAA requirements.
Under The Hood
Architecture
- Monorepo structure using pnpm workspaces to cleanly separate desktop, web, and native components, enabling independent development and deployment cycles
- Next.js frontend organized with clear layers for UI, API routes, and authentication, supported by strict environment management and containerized deployment patterns
- Backend services decoupled via Docker Compose with explicit boundaries between media processing, database, and storage layers, communicating through REST and webhooks
- Native Rust components integrated via Cargo workspaces to handle performance-critical tasks like video parsing and frame extraction, with well-defined FFI boundaries to JavaScript
- Infrastructure-as-code approach with orchestrated Docker services, health checks, and environment templating, eliminating implicit dependencies
Tech Stack
- Next.js 14 and TypeScript for the web application, powered by Turborepo for monorepo build orchestration
- Rust-based desktop and media server components using Tauri, with custom forks of wgpu, cidre, and nokhwa for low-level audio/video capture and rendering
- MySQL 8.0 managed via Prisma ORM for type-safe schema evolution, deployed with automated migrations and health monitoring
- MinIO as S3-compatible storage with automated bucket provisioning via CLI tools in containerized environments
- Solid.js and @kobalte for reactive desktop UI, paired with Vitest and Biome for testing and code quality enforcement
- Multi-environment Docker deployments with fully orchestrated services and environment injection for self-hosted production readiness
Code Quality
- Comprehensive test coverage spanning unit, integration, and end-to-end scenarios, with mocked dependencies and real-world system validations
- Strong type safety enforced through Effect and custom type guards, modeling domain entities with invariants encoded at the type level
- Robust error handling with structured boundaries and custom error types, ensuring predictable failure modes across frontend and backend
- Consistent, intent-driven naming conventions that align function names with domain responsibilities and improve maintainability
- Automated linting, test scaffolding, and schema generation for SEO metadata, ensuring consistency between UI and structured data outputs
What Makes It Unique
- Seamless cross-platform video editing with synchronized timeline editing between desktop and web, eliminating external sync dependencies
- Intelligent mask track system that auto-positions clip segments using dynamic spatial analysis, reducing manual editing overhead
- Application-specific drag-and-drop protocol (application/cap) that preserves metadata when moving video files between interfaces
- Stripe webhook handler that auto-creates guest accounts from payment events, removing friction in user onboarding
- Lexical-style editor extensions that treat video overlays as first-class editable objects, not static layers