Unlimited video with recordings that live in your Drive, live transcription on every call, and per-listener translation in 12 languages — plus agents that join with a real voice and face.
Live product Video means unlimited minutes and recordings that land in your own Drive — not locked in a vendor cloud behind an upgrade. Every call is indexed by the Company Brain, so a meeting from last quarter is searchable months later by what was actually said.
Live product Before anyone sees you, the Green Room is a full pre-join lobby on a 16:9 self-preview. Pick your camera, microphone and speaker from enumerated devices, watch a live mic-level meter, and play a one-tap speaker test tone. Flip on blur or a virtual background, and the one-tap Auto-light, Auto touch-up, Auto-frame and Noise-suppress buttons. An Advanced panel adds a lighting slider from 70% to 150%, join-muted and join-without-video defaults, captions-on by default, and your preferred layout. A live strip shows who is already inside, and a host can turn on manual guest admission — a waiting room — right from here.
Live product Plantel Translate is per-listener and OFF by default. Each participant turns it on for themselves — the speaker does nothing. With it on, a speaker in a different language is auto-translated into the listener’s language with captions and a natural voice while the original audio is ducked; a same-language speaker passes through normally with no captions. It runs on the free in-browser voice engine, so there’s no API, AI or usage cost, and any failure degrades to captions-only so the call never drops.
Live product Virtual backgrounds use genuine on-device person segmentation — your camera frame is matted per-pixel against a green-screen-free background, not a lazy box blur — and a matte-refinement pass kills the cloudy halo while keeping hair intact. Pick blur or one of 24 backgrounds (18 photographic scenes plus 6 gradients) in the Green Room. Audio runs through deep-learning noise suppression on the mic everyone else hears, so keyboard clatter, music and the table next to you drop out for the whole room — and your raw mic publishes instantly, then hot-swaps to the cleaned track, so a slow load can never delay or mute you.
Live product When a Plantel agent joins a Huddle it appears camera-on by default as a live, rigged 3D body wearing its real face — idling with a breath, looking at the camera, talking with mouth movement, and performing gestures on command. Type or say “raise your hand,” “wave,” “nod” or “thumbs up” and it does it. It replies in the room with persona-aware text in chat, on the caption strip, and spoken aloud, with the same audio driving its face. The renderer is throttled and pauses when its tile is hidden, degrading to a clean headshot before it could ever tax the call.
Live product Agents speak with Plantel Voice: a voice engine with a deterministic per-agent identity built from a seed (gender, age, pitch, rate, accent), so the same agent always sounds like itself across every call, with effectively unlimited distinct voices. Smart-tone prosody reads the text and shifts delivery — slower and softer for empathy, brighter and faster for excitement, deeper and measured for serious moments — auto-detected from cue words. The same audio stream lip-syncs the agent’s on-screen face.
Live product Write a script — # for section headings, (pause) for cues — saved to a reusable, Brain-backed library. The teleprompter is draggable, resizable, and tunable for opacity, contrast, font size and line height, with a mirror mode. Turn on Follow my speech and it indexes your script and uses live speech recognition to scroll to exactly where you are, auto-pausing when you stop talking. Present a deck and the current slide is published to the main stage on the same path as a screen share, so every participant sees your slides — with a presenter view showing the next slide, a timer and arrows. The script and deck stay two-way linked.
Live product Recording captures real video that saves straight to your Drive — not a vendor cloud — and the Huddle hub keeps Home, Upcoming, Recordings and History tabs with live counts of what’s live, upcoming and recorded. After you stop, the Studio editor turns the take into an asset: an AI-generated title and summary, an instant shareable link, trim with in/out points, chapters, captions and transcript, an end call-to-action, a camera-bubble position, and viewer analytics — built to beat a single-purpose screen recorder, with the file living in your own pool.
Live product Hosts get owner-tier moderation: mute anyone, mute everyone, pin or unpin a tile to the stage, remove someone, lock the room to block new joins, admit or deny from a live waiting list, and end the huddle for all. One rule is built in by design — a host can mute a human and can mute or unmute an agent, but can never force-unmute a person. That arrives as a polite “Ask to unmute” nudge, so the human always controls their own mic. Identity — name, role, company, guest flag — comes straight from the call token, so every tile is trustworthy.
Live product Plantel AI sits in the room, follows the conversation off the live transcript, and at the end produces a recap — summary, decisions, per-person talk-time split, an engagement score, and the action items it heard. One click files those action items straight into HQ Tasks for the right people, and the recording itself is indexed by the Company Brain so the meeting is searchable later by content, not just title. The end screen even counts down 60 seconds to your next huddle.
Live product No bolt-on apps, no recorder bot in the corner. The whole engagement toolkit ships in the call.
Each speaker’s mic is transcribed live and broadcast to the room, attributed by name — humans by identity, agents by agent name. Finals accumulate into a running transcript that feeds the engagement read and the end-of-call recap. No recorder bot to invite.
A live read built from real signals, never a random number: talk-time balance, participation, cameras-on, activity and caption sentiment, multiplied by a local attention factor. Banded Excellent to Low, with a click-through breakdown that carries into the recap.
Eight floating reactions broadcast to the room, and raise-hand puts you in a numbered queue with a ding and a “you’re #N” toast. The host sees the ordered list and can lower hands — so the room stays orderly.
Launch a poll with comma-separated options and watch results render as live percentage bars. Q&A lets anyone submit a question and upvote others, so the best questions rise to the top. Both feed the engagement score.
Split a big call into named breakout rooms — each with an icon, an occupant list and a capacity — then pull everyone back with a tap. No separate tool, no new meeting link.
A real canvas whiteboard built into the call, alongside captions, polls and breakouts, so you can sketch the idea live without launching another app.
Share a screen, window or tab at full frame rate with audio; it lands on a dedicated content pane for everyone, not a cramped tile. The browser’s own Stop bar routes to the same clean unshare — no awkward reconnect, no dropped call.
Switch between Auto, Gallery, Speaker and focus layouts on the fly. In spotlight mode the stage auto-follows the loudest active speaker, or pin any tile to keep it center stage.
Pull the right teammate — or the right agent — straight from the in-call directory, searchable by name, role or department, with live presence dots telling you who’s free. Or copy an invite link or email a guest.
One-tap prompts off the live transcript: Summarize so far, Action items, Decisions. The answers come from what was actually said in the room, so you leave with the notes already written.
A live connection chip reads your link and labels it Excellent, Good, Fair or Offline, next to a running call timer — so you always know where you stand before the meeting matters most.
Drive the whole call from the keyboard: M mute, V camera, S share, R raise hand, C captions, L layout, P people, T chat, A AI assistant. Stateful glass-bubble controls turn red when off and teal when on, in both light and dark themes.
Unlimited video, owned recordings, and translation for everyone — in the seat.
