Videorc is an open-source, AI-native desktop studio for creators: record your screen and camera, stream to multiple platforms at once, and bring Orcle, an AI producer that keeps up with your chat while you're live. All from one window.
Current release tracks
- macOS Beta → — macOS 13+, Apple Silicon. Signed and notarized beta builds are the current public desktop release.
- Windows Alpha guide → — Windows 11 x64 only. A Windows installer is published only after the signed candidate, clean-machine, malware-scan, update, uninstall, and real-device acceptance gates pass. A CI artifact or unsigned local package is not a public alpha.
Both tracks are pre-release software. Expect fast-moving releases, rough edges, and occasional recording or streaming bugs while the app is being hardened.
Most capture tools make you choose between "simple but shallow" and "powerful but a cockpit". Videorc aims for the third option: a studio that is genuinely simple to run — pick a scene, hit record — while the heavy lifting (a native capture engine, multi-platform streaming, live captions, an AI producer) happens underneath.
- Scenes, not knobs. Screen + camera, screen only, camera only, or side-by-side splits — with draggable camera placement, corner snapping, shapes, and framing controls.
- Backgrounds with taste. Bring your own wallpaper (PNG/WebP/JPEG), tune its visibility with one slider, or remove it for a full-bleed recording.
- Record and stream in one pipeline. Local MKV recording (with automatic MP4 remux), RTMP streaming, or both from a single encode — including simulcast fan-out to multiple destinations with per-target health status. Streaming quality is free on every plan: up to 1080p60, and true 4K30 on YouTube (other platforms ingest up to 1080p).
- Live captions. Streaming speech-to-text (~1s latency) with optional caption burn-in on the stream, the recording, both, or neither.
- Orcle, your AI producer. One switch, "Orcle joins my streams", and Orcle catches questions from every platform with a drafted reply you approve, flags spam and abuse against your own rules, greets first-timers and reminds you of your promises. It never posts or acts on its own. Cloud AI is opt-in, and after each stream a short report stays on your computer. See docs/orcle-live.md.
- Platform-aware preview. macOS uses a detached native CAMetalLayer preview. Windows Alpha currently uses the documented uncompressed, latest-wins Electron proof surface; it must not be described as CAMetalLayer or as final native parity.
- Gated auto-updates. Signed, notarized macOS Beta builds update in place. Windows updating is a release-candidate gate and is enabled only for an accepted, signed Alpha feed.
- Electron + React desktop shell (TypeScript, shadcn/ui) for the studio UI.
- Rust backend owns capture, composition, recording, and streaming; the shell talks to it over an authenticated localhost WebSocket protocol.
- FFmpeg (an LGPL-compliant build, bundled) drives encoding and output.
- SQLite local session library — your recordings and Orcle's stream reports stay on your machine.
The desktop app — capture, scenes, recording, streaming, captions UI — is free software under AGPL-3.0. You can build it, run it, and audit every line that touches your camera, microphone, and screen.
Cloud AI features (live captions and Orcle) run through a signed-in Videorc account: the desktop app never holds AI provider keys, and nothing goes to the cloud until you turn it on. Hosted AI is what funds the project.
Prerequisites: Node.js 24.x and Rust stable (rustup). macOS source development
uses FFmpeg on PATH; Windows development fetches the pinned LGPL bundle with
pnpm ffmpeg:fetch:windows. The checked-in .node-version, engines, and
packageManager fields keep local tools aligned with CI; Corepack installs the
pinned pnpm 11.
corepack enable
corepack install
pnpm install
pnpm devThe app launches the Rust backend automatically. On macOS, recordings default
to ~/Movies/Videorc/Recordings and session metadata lives in
~/Library/Application Support/Videorc/videorc.sqlite3; Windows uses its native
user data and Videos locations.
Developing on Windows? See docs/windows-dev-loop.md for setup, the version-floor escape hatch, and the fast verify loop.
To produce a local unsigned macOS app bundle:
pnpm ffmpeg:build:macos # build or reuse the bundled LGPL FFmpeg
pnpm package:desktopSee docs/distribution.md for signing, notarization, and FFmpeg distribution details.
AGENTS.md is the contributor guide: verification gates, recording and native-preview rules, and repo conventions. The short loop:
pnpm typecheck
pnpm build
pnpm smoke:dev # records a test-pattern MKV per layout preset — no permissions needed
cargo fmt --check
cargo test
cargo clippy -- -D warningsThe full non-packaged acceptance gate (what CI runs) is:
pnpm smoke:local-gatesNotable smokes: pnpm smoke:multistream proves simulcast fan-out end to end
against local RTMP listeners (including the offline-destination failure
guarantee), and pnpm smoke:packaged exercises a packaged build. None of the
default cross-platform smokes require camera, microphone, or screen permissions.
The Windows installer lane additionally runs pnpm smoke:windows-native-screen
against DXGI (gdigrab fallback) and keeps the packaged proof preview live during
a real ScreenOnly recording.
Videorc's macOS track is in Beta and its gated Windows track is in Alpha. Bug reports with reproduction steps are very welcome; Windows testers should use the privacy-safe Windows Alpha issue form and never attach credentials, recordings, or an unverified support bundle to a public issue. For larger changes, please open an issue first so we can agree on the shape before you invest in a PR. Read AGENTS.md before touching recording or native-preview code — those areas have non-negotiable verification gates.
Code: AGPL-3.0. Brand: the Videorc name, logo, and app icon are not part of the code license — see TRADEMARK.md before distributing a modified build. The bundled FFmpeg is built LGPL-compliant; see docs/distribution.md for third-party licensing notes.