VIDEOBOX Documentation

A complete TV studio inside your computer: it records, directs itself by who is speaking, streams to every network at once and runs video meetings directly — no cloud, no server subscription, no video editor.

Until today, professional video required five tools and three people: recording software, streaming software, Zoom for guests, a video editor — and a director deciding who is in frame. VIDEOBOX unites them all in one screen and replaces the director with an algorithm.

What used to be complicated — now happens by itself:

• Automatic routing: the stage moves to whoever is speaking, like a newsroom. You speak — the camera cuts to you.
• What you see is what gets recorded: the designed stage, titles, guests and mixer go straight into the recording and the stream — no "we'll fix it in editing".
• Recording, live streaming and meetings — one screen: switch mode with one click, all the design stays.
• AI that runs on your computer: transcription, background removal, face tracking and noise removal — without sending your video to any server.
• No subscriptions in between: direct P2P meetings, direct streaming to networks, files on your computer.

The result: a single creator today produces what used to require a studio — a four-participant webinar, a video podcast, a live stream to five networks and vertical reels — in the same app, in the same session.

Two or more on stage? VIDEOBOX listens to each participant separately and moves the frame to whoever is speaking — in real time, both in recording and live. It's the first time a single creator gets multi-camera directing without touching anything.

How it works:

1. Each participant — local camera, wireless phone or remote guest — gets its own audio channel.
2. A speech detector measures all channels 10 times per second and identifies who is speaking continuously.
3. The stage cuts to the speaker within ~0.3 seconds, stays at least one second, and returns to a wide shot after silence.
4. A short "yes, right" doesn't trigger a cut; when several speak together — a wide shot of everyone.

Two directing styles, by stage:

Full control in the host's hands:

• Response speed: fast (TV), balanced, or relaxed (podcast).
• Microphone sensitivity: low / medium / high — for quiet and noisy rooms.
• Click to lock: clicking a participant on stage or 📌 in the list locks on them — even mid-recording. Another click releases.
• Automatic chapters: every switch between speakers is recorded as a chapter — viewers jump straight to the speaker they care about.
• Speaking indicator: a green halo on whoever is speaking, in the preview only — never in the file.

Routing recognizes phones, meeting guests and the main camera. Several local cameras share one microphone — so to direct between two people in the same room, the second participant's phone is the second camera (and comes with its own microphone).

All the AI in VIDEOBOX runs locally, on your CPU and graphics card. Video and audio never leave the computer, there are no usage minutes and no monthly subscription — and it works even without internet.

The models (Whisper, MediaPipe, RNNoise) are bundled with the app or downloaded once. Transcription lets you choose between a fast model and a more accurate one, depending on your computer.

What you see on stage is exactly what goes into the file and the stream. Background, frame, titles, logo and effects — all burned in real time, with no editing after recording.

• 28 designed stages in one click: news, executive, podcast, neon, magazine, gamer, luxury and more — for one or two people, landscape (16:9) and vertical (9:16) for Reels, TikTok and Shorts.
• 20 live stage backgrounds and animated frames around the camera and the stage, in your brand colors.

• Animated titles engine: lower third, full bar, running ticker and call-to-action button; entry, exit, speed, display time and auto-repeat. A phone number running on screen at the press of a button — mid-recording or stream.

• Scenes and transitions: intro, camera, screen + camera, "be right back" — one-click transition with fade, slide, zoom, wipe or a transition video. Studio mode prepares the next scene without it going on air.
• Free layers: text, image, video, clock, shape and QR code — drag, rotate and resize directly on the stage.

• Designed screen share: screen + camera bubble, screen + camera column, or screen only; a floating control panel that doesn't enter the recording.

• Branding: logo / watermark, saved branding profiles, and any font from Google Fonts.

• Teleprompter and agenda: a scrolling script or topic list beside the stage — for the host only; moving to the next item creates a chapter in the video.
• Replay: save the last seconds as a clip — even without recording running.

Same stage, same titles and the same automatic routing — one click on "Go live" and everything goes out to the world.

• Multistreaming: YouTube, Facebook, Instagram, Twitch, LinkedIn, X and any RTMP server — in parallel, one encoder.
• Connection check before every stream: DNS, TCP, TLS and RTMP handshake for each network; a network that fails doesn't bring down the others. A test stream measures the actual bitrate.

• Recording alongside the stream — the local file in full quality, independent of internet quality.
• Comments from all platforms in one place: YouTube, Twitch, Facebook and Instagram — with per-network filtering; clicking a comment pops it onto the stage, or a running chat burned into the stream.

• Streaming an existing recording: pick a video from the library and stream it through the studio — with live titles and real-time comments.
• Stream health: bitrate, FPS, speed and dropped frames — per network, with early warning of a slow connection.
• Quality by connection: 720p up to 1080p at 8 Mbps, with automatic calculation of the required upload.

A meeting in VIDEOBOX is a meeting that takes place on a TV stage. Video and audio travel directly between computers (P2P), end-to-end encrypted — there is no media server in the middle that sees or hears the call. The shortest path between participants — and therefore the fastest.

• One link and that's it: participants join from the browser, on computer or mobile, without installing anything. The invitation also includes an app download link and a QR code; send via WhatsApp or email in one click.
• Participants on stage: every guest is a camera source and a mixer channel — participant grid, titles, automatic routing to the speaker, recording and live streaming of the meeting itself.
• What participants see of you: camera, shared screen — or the whole designed stage with titles.
• Full management: lock meeting, remove participant, per-participant volume, who's on stage, chat, raise hand and connection quality indicator. Participants see when the meeting is recorded.
• No participant limit: quality adapts to room size (720p up to four, then gradually lower). On a home connection up to 6–8 participants is recommended — the app warns in advance.
• Introduction server only, free: a small server on Cloudflare (free plan) only introduces participants and serves the join page; an optional encrypted relay for blocked corporate networks.

Up to 25 participants on stage, from three kinds of sources — and they all behave the same: picture to the stage, sound to the mixer, routing to the speaker.

• Placement templates: split, cards, picture-in-picture, focus, interview and meeting grid — arrange themselves for any number of participants, landscape and vertical. Participant names are burned onto the stage.
• Professional audio mixer: a channel per source with fader, low-cut filter, noise removal, noise gate, compressor, A/V sync and headphone monitoring; presets for voice, podcast and noisy room. A master limiter prevents distortion.

• Multi-track recording: each source is also saved separately (MKV) — for professional podcast editing after recording.
• Smart mute: a phone is heard only when it's on stage or selected as a microphone — no echo and no surprises.

The recording ends — and opens straight in the editor, with chapters created automatically from every speaker switch, scene or agenda item. Any video file from your computer (MP4, WebM, MOV, MKV) opens in the same editor.

• Multi-track timeline: video, titles and subtitles — drag, trim, split, with a transition at every cut point.
• Animated titles and designed subtitles: color and font sets per project, subtitle styles, SRT/VTT import and export.
• A format for every network: 16:9, 9:16, 1:1, 4:5 — with blurred background, center crop or branded stage; image tone (contrast, saturation, brightness, temperature).
• One-click pro tools: AI transcription, audio enhancement, silence cutting, speed-up with pitch correction, and instant cutting without re-encoding.
• MP4 or WebM export — the result is saved as a new recording in the library; the source file is untouched.
• Local library: all recordings with thumbnails, search, automatic saving to the Downloads folder, and live streaming of any recording in one click.

One screen for recording, streaming and meetings — switch mode with one click, and the sidebars arrange themselves by mode.

• Simple / Full: start with only the main tools; one click reveals mixer, scenes, layers, camera enhancement, shortcuts and technical stats.
• Customizable sidebars: two sides with collapsible areas — choose what's shown, on which side and in what order; each sidebar collapses to an icon strip and the stage expands. Everything is saved.
• Light / dark on every screen, full Hebrew and English with correct direction.
• Stages by format: in 16:9 you only see landscape stages, in 9:16 only vertical ones — no confusion.
• Global keyboard shortcuts: recording, mute, scenes and titles — even when the studio is in the background.

Example: a BOX CRM webinar — host on camera, two guests from the browser, a "news" stage, automatic routing to the speaker, a ticker with a phone number, streaming to YouTube and Facebook in parallel, comments from both networks on screen, and a recording with chapters uploaded to the website the same evening — one person, one computer.

Everything stays with you: recordings are saved on your computer, the AI runs on your computer, meetings travel directly between participants, and the stream goes straight to the networks.

• No account, no login, no monthly subscription.
• No cloud upload — not even for transcription or background removal.
• The studio works even without internet; only live streaming and meetings need a connection.
• One Windows app — one-click install, Hebrew from the first moment.

VIDEOBOX is part of the BOX family by theouterspace.online.