> ## Documentation Index
> Fetch the complete documentation index at: https://docs.trailercast.io/llms.txt
> Use this file to discover all available pages before exploring further.

# Recording demos

> Three ways to get demos into TrailerCast.

<Tabs>
  <Tab title="Upload existing recording">
    The simplest path. If you've already recorded a demo with Zoom, Loom, Riverside, etc., you can upload it directly.

    <Steps>
      <Step title="Click Upload demo">
        Open the **Library** from the left nav → **Upload demo** button (top right of the Library page).
      </Step>

      <Step title="Drag-and-drop your file">
        MP4 is recommended. MOV, M4V, and most browser-recordable formats also work.
      </Step>

      <Step title="Optional: enter prospect name">
        If you don't, we'll use the filename.
      </Step>

      <Step title="Click Upload">
        Watch the progress bar.
      </Step>
    </Steps>

    ### Processing stages

    | Progress | What's happening                                      |
    | -------- | ----------------------------------------------------- |
    | 5–15%    | Uploading to our servers                              |
    | 15–35%   | Transcribing with Deepgram                            |
    | 35–50%   | Cleaning the transcript (removing umms, false starts) |
    | 50–65%   | Detecting valuable moments                            |
    | 65–90%   | Composing the trailer                                 |
    | 90–100%  | Uploading the final trailer                           |

    When it hits 100%, you're auto-redirected to the trailer review page.

    ### File requirements

    * **Length**: 5 minutes minimum, 90 minutes maximum
    * **Sweet spot**: 15–45 minutes
    * **Size**: under 2GB (split into two recordings if larger)
    * **Audio**: clear, ideally per-speaker tracks if available
  </Tab>

  <Tab title="Calendar bot (auto-schedule)">
    If you connect your Google or Microsoft calendar, TrailerCast can automatically schedule a recording bot on every meeting that matches your filter rules.

    <Steps>
      <Step title="Connect your calendar">
        See [Integrations](/setup/integrations#calendar).
      </Step>

      <Step title="Toggle auto-scheduling on">
        By default, the auto-scheduler looks for meetings whose title contains **"demo"** or **"trailercast"**. Edit the keyword list to taste.
      </Step>

      <Step title="System polls every 10 minutes">
        Bots get scheduled on matching events. View them in **Calls → Upcoming**.
      </Step>

      <Step title="On the day">
        The bot joins automatically as a participant. After the meeting ends, the recording flows through the same pipeline as an uploaded file.
      </Step>
    </Steps>

    <Tip>
      Bots are visible to all attendees, meetings are explicitly recorded with attendee awareness, never silently. We use Recall.ai for the bot infrastructure.
    </Tip>
  </Tab>

  <Tab title="Record in the browser">
    Record a screen + camera trailer directly in TrailerCast — no Zoom, no Loom, no upload step. Lands as a normal trailer ready to share or attach to Decision Rooms.

    <Steps>
      <Step title="Open the recorder">
        From **Trailers** → the **Reusable** tab → **Record**. It's the primary button on that tab, because a reusable clip is something you record rather than something a call produced.

        On the **Deal** tab, recording still lives under **+ New trailer ▾** → **Record in the browser**.

        The recorder opens in a **separate browser tab**, not a modal, so a screen share never captures TrailerCast's own interface and you can drag the recorder onto a second monitor. If your browser blocks the new tab, TrailerCast says so and offers **Open recorder here** to load it in the current tab instead.
      </Step>

      <Step title="Pick a layout">
        * **Screen + camera (PiP)** — screen recording with your webcam in a circle, bottom-right corner.
        * **Screen only** — just the screen, no webcam.
        * **Camera only** — just your webcam, no screen.

        Mic is on by default. Toggle off if you want a silent walkthrough.
      </Step>

      <Step title="Click Start recording">
        Your browser asks for screen + camera + mic permissions. Approve them. The recording starts; a red dot + timer shows in the modal, and your browser's omnibox shows a "Sharing your screen" indicator.
      </Step>

      <Step title="Pause / resume / stop">
        Up to 15 minutes per recording. We show a soft warning at 10 minutes ("getting long, want to save and start a new clip?") and auto-stop at 15.
      </Step>

      <Step title="Preview, then save">
        Play it back. If you fluffed it, **Discard + record again**. If it's good, **Save as trailer**. The recording uploads to your trailers and appears in the list immediately — no AI processing wait.
      </Step>
    </Steps>

    ### What's different from uploaded recordings

    Browser-recorded trailers **skip the AI pipeline**. The recording is treated as the final trailer — no transcription, no moment detection, no auto-built highlight reel. Use this when:

    * You already know exactly what you want to record (60-90 seconds is the sweet spot)
    * You want to ship NOW, not wait 5-10 minutes for AI processing
    * The recording is a personalized intro / nurture clip / async update, not a long demo
    * You want a polished short clip you can attach to Decision Rooms or send via per-recipient share links

    For long-form demo walkthroughs that benefit from AI moment detection (the auto-built 90-second highlight reel), upload them through the **Upload existing recording** flow instead.

    Browser recordings made without a prospect attached land on the **Reusable** tab of the Trailers screen, since that's what they usually are: a clip you'll send to many people. If one turns out to belong to a single deal after all, move it with **Move to Deal** in the trailer's ⋮ menu — and the same menu moves a deal trailer the other way when it becomes something you want to reuse.

    ### Browser support

    Works on Chrome, Edge, Firefox, and Safari on desktop. **iOS Safari does NOT support screen capture** — Apple blocks it. The modal shows a "desktop only" notice on iPhone/iPad.

    ### Coming soon

    Per-prospect personalization layers (record a short webcam intro stuck on top of any existing trailer, one per stakeholder), branded intro/outro slide auto-composition for recorded clips, and playlist composition (multi-clip stitched sequence with AI narration on transitions). All designed; build incrementally based on usage.
  </Tab>

  <Tab title="Manual bot (one-off)">
    For one-off calls that don't match your auto-schedule rules, say a partner sync you want recorded for review.

    <Steps>
      <Step title="Calls → Upcoming → Schedule new bot">
        From the Calls phase.
      </Step>

      <Step title="Paste the meeting URL">
        Zoom, Google Meet, Microsoft Teams.
      </Step>

      <Step title="Choose recording kind">
        * **Demo**, full demo pipeline (transcribe + moments + trailer)
        * **Qualifying Call**, transcribe + AI summary + topics, no trailer
        * **Meeting**, generic transcribe + AI summary, no trailer
      </Step>

      <Step title="The bot joins at the scheduled time">
        Same flow as auto-scheduled bots from there.
      </Step>
    </Steps>

    <Tip>
      Misclassified after the fact? You can [reclassify](/calls#reclassifying-a-call) any time after the recording completes, it re-processes through the appropriate pipeline.
    </Tip>
  </Tab>
</Tabs>

## What's NOT a demo

A few non-obvious things that won't process well:

* **Slide-only meetings with no narration**, the AI infers "live demo" from spoken language, not pixels. A 30-min slide deck without verbal walkthrough produces a near-empty trailer.
* **Mostly small-talk meetings**, the AI explicitly filters out chitchat. If your call is 80% rapport-building, there won't be much trailer material.
* **Bad audio**, choppy / muted / heavily-compressed audio causes transcription errors that cascade through the pipeline. Use a real microphone.

## Three recording kinds explained

| Kind                | Pipeline | Output                                                                                         |
| ------------------- | -------- | ---------------------------------------------------------------------------------------------- |
| **Demo**            | Full     | Trailer + transcript + moments + summary + deal brief                                          |
| **Qualifying Call** | Lite     | Transcript + topic-extracted summary, no trailer (turn it into a [Buyer Space](/buyer-spaces)) |
| **Meeting**         | Generic  | Transcript + key takeaways + action items, no trailer                                          |

Pick the right kind at scheduling time when possible, it determines which pipeline runs and what gets generated.