Launchscene
PricingSign in
Explore

Introducing Fish Agent

Fish Audio · Nov 2026 · 2:37 · AI-annotated

Length
2:37
Scenes
18 · 8.8 s avg
Product on screen
0:00
Features shown
2
Call to action
2:09 · 82%
Voiceover
Yes
00:00:00 / 02:37:17

Fish Audio’s 2:37 launch film runs 18 scenes, about 8.8 s each. The product is on screen from the first frame and 2 features are revealed. The call to action lands at 2:09 (82% in), over a voiceover. A single presenter anchors every live-action beat in a cozy brick-walled room, while frosted-glass panels, pop-in icon chips, and dotted-line diagrams explain the architecture beside him. The pacing is steady and conversational, built around one long uninterrupted transcript demo, and the video closes by turning the social post itself into the call to action.

Footage neededEstimated needs: 2 capture groups — your screens or recordings · 1 official logo — reused across the film · 12 filmed passages — shoot, supply or replace · 2 3D passages — build or supply

Is this your launch? to your site or press kit.

More like this

Introducing Fish Agent, scene by scene

· 18 scenes as text
  1. 01 · Product intro · 0:00–0:04

    A man in an olive sweater sits at a wooden table against a brick wall, gesturing, a laptop beside him showing the Fish Audio logo. On the left, a waveform icon and the title 'Live 3' sit above the subline 'Full duplex voice model', which builds word by word.

    On-screen text: “Live 3”, “Today we're launching Live 3, a full duplex”, “Full duplex voice model”, “model”, “And the demo is a little crazy.”

    Techniques: Kinetic type

    Camera: Static

  2. 02 · Product intro · 0:04–0:07

    A man in a brown sweater sits before a brick wall, speaking to camera, a laptop with a waveform at the right edge. A lower third with his name and title fades in lower left, subtitles caption his words at the bottom. Locked-off framing.

    On-screen text: “Cale Shapera”, “Founding Forward Deployed Engineer | Fish Audio”, “You're about to hear two AI agents on”, “a phone call.”

    Camera: Static

  3. 03 · Demo · 0:07–1:23

    A man at a desk delegates a blender complaint to his laptop; the shot reframes to a close-up of the screen, where a support-call transcript builds line by line beside a floating card for the AI agent Maya, while feature labels pop in at the top right.

    On-screen text: “Cale Shapera”, “Founding Forward Deployed Engineer | Fish Audio”, “So Fish, that fancy blender you ordered for me stopped working.”, “Can you call support and sort it out?”, “Sorry, I will take care of it.”, “Natural, expressive voice”, “Context-Aware Intelligence”, “Emotion-aware responses”, “Adaptive negotiation”, “Customer retention”

    Techniques: UI zoom, Split screen

    Camera: Push in

  4. 04 · Demo · 1:23–1:29

    A man in a brown sweater sits before a laptop against a brick wall, head down as he speaks. The laptop screen shows a transcript with emotion tags and a waveform. White subtitles on a dark bar appear at the bottom. The camera holds a locked-off medium shot.

    On-screen text: “Thanks, Fish.”, “And my car's actually been making a weird”, “noise.”, “Can you call some mechanics and book us”, “in for this week?”

    Camera: Static

  5. 05 · Problem · 1:29–1:38

    A man in a brown sweater sits at a wooden table against a brick wall, talking and gesturing to camera. A frosted glass panel slides in at left, and four icon tiles (mic, T, brain, speaker) pop in one by one, linked by dotted lines. Subtitles run along the bottom.

    On-screen text: “Well, that's Live 3 for you.”, “Typical voice platforms turn your voice into text”, “Typical Voice Platform”, “Timing relies on separate turn taking models.”, “Speech Recognition”, “STT”, “and LLM writes a reply and then another”, “LLM”, “TTS”, “model turns it back into audio.”

    Camera: Static

  6. 06 · Problem · 1:38–1:43

    A man in a brown sweater talks to camera before a brick wall. A rounded card titled 'Typical Voice Platform' overlays the left with icon tiles linked by dotted lines; a clock tile labeled 'Turn-taking model' pops in between STT and LLM, widening the row. Subtitles run along the bottom.

    On-screen text: “Typical Voice Platform”, “Timing relies on separate turn taking models.”, “Speech Recognition”, “STT”, “LLM”, “TTS”, “Turn-taking model”, “And then a separate model keeps guessing when”, “you've stopped talking.”, “That's why talking to them feels unnatural.”

    Camera: Static

  7. 07 · Problem · 1:43–1:46

    A man in a brown sweater sits at a wooden table against a brick wall, talking and gesturing to camera. A translucent pill-shaped input box fades in at left and types a short prompt, while a caption line runs along the bottom.

    On-screen text: “And it's an architectural problem, so you can't”, “Typical Voice Platform”, “prompt your way out of it.”

    Camera: Static

  8. 08 · Feature reveal · 1:46–1:53

    A man in a brown sweater sits in a medium close-up before a brick wall and talks. On the left, a circular Live 3 badge with a waveform builds, linked by wavy lines to microphone and speaker icons labeled Listens and Speaks, with Manages turn taking fading in below. Subtitles run at the bottom.

    On-screen text: “Live 3 solves this with a full duplex”, “Live 3”, “model that listens, speaks, and manages turn-taking”, “Listens”, “Speaks”, “Manages turn taking”, “simultaneously, keeping the conversation flowing naturally and responding”

    Camera: Static

  9. 09 · Feature reveal · 1:53–1:57

    A man in a brown sweater sits at a wooden table against a brick wall, gesturing with open hands as he speaks to camera. White subtitles run in a dark bar along the bottom of the locked-off frame.

    On-screen text: “Live 3”, “Listens”, “Speaks”, “Manages turn taking”, “simultaneously, keeping the conversation flowing naturally and responding”, “in context.”

    Camera: Static

  10. 10 · Feature reveal · 1:57–2:04

    A man in a brown sweater speaks to camera in a warm room with a brick wall, lamp, and plant behind him. White captions track his words at the bottom, while a glowing “Tools” chip with an orange icon fades in on the left as he gestures.

    On-screen text: “Starting today, you can access Live 3 on”, “Fish Agent Platform, our no-code voice agent”, “builder with built-in tools, knowledge, and everything”

    Camera: Static

  11. 11 · Feature reveal · 2:04–2:09

    A man in a brown sweater sits at a wood table before a brick wall, talking and gesturing. Dark rounded pills with orange icons (Knowledge base, Tools, Guardrails, Analysis, Tests) pop in around him along a dashed circle as subtitles roll across the bottom.

    On-screen text: “builder with built-in tools, knowledge, and everything”, “you need to build agents that can both”, “make and answer calls.”

    Camera: Static

  12. 12 · CTA · 2:09–2:11

    A man in a brown sweater sits in front of a brick wall, a lamp and plant behind him, speaking to camera. The frame is locked off as he talks, with white subtitles at the bottom.

    On-screen text: “If you're an enterprise, go to the link”, “in the caption, paste your website, and get”

    Camera: Static

  13. 13 · CTA · 2:11–2:20

    A browser window of the Fish Audio site sits on a peach gradient, zooming in on a two-card demo. The left card's teal avatar relabels from Your agent to Tesla support while a red call timer counts up; the right card swaps its email prompt for a ready message.

    On-screen text: “in the caption, paste your website, and get”, “a custom voice agent built for your company”, “instantly.”, “Hi, thanks for calling Tesla.”, “This is Clara.”, “How can I help you today?”, “And as part of this launch, we're giving”

    Techniques: UI zoom

    Camera: Push in

  14. 14 · CTA · 2:20–2:26

    A man in a brown sweater sits at a wooden table before a brick wall, speaking to camera and gesturing with both hands. A lamp, plant and framed print fill the background. Subtitles swap at the bottom edge.

    On-screen text: “And as part of this launch, we're giving”, “away free credits to build voice agents and”, “a free number to go live.”

    Camera: Pull out

  15. 15 · CTA · 2:26–2:27

    A man in a brown sweater speaks in a post video that fills the frame, its caption reading 'a free number to go live.' then 'Just like, save, repost, and comment Live 3.' The view pulls back to show the full card with action icons and a comment thread.

    Techniques: UI zoom

    Camera: Pull out

  16. 16 · CTA · 2:27–2:33

    A social post interface shows a man speaking in its video player while a reply box types “Live 3”. The view zooms into the video, the man stands and exits frame right, and a white bar waveform animates over the empty chair.

    On-screen text: “And we'll send you the link.”

    Techniques: UI zoom

    Camera: Push in

  17. 17 · CTA · 2:33–2:35

    A phone lies angled on a wooden table, its lock screen showing the time and a notification. On the left, a white waveform logo, a “Book a demo today” button, and the fish.audio wordmark fade in over the wood.

    On-screen text: “Book a demo today”, “fish.audio”

    Techniques: 3D mockup

    Camera: Static

  18. 18 · CTA · 2:35–2:37

    A 3D iPhone lies at an angle on a wooden table, its lock screen showing the time and a notification. To the left, a white sound-wave logo animates its bars above a Book a demo today button and the fish.audio wordmark, all held on a locked frame.

    On-screen text: “Book a demo today”, “fish.audio”

    Techniques: 3D mockup, Morph

    Camera: Static