Botverse Conform

What "conform" means here

A conform is what a post-production pipeline calls the step where a picture track and a separately-sourced audio track get combined into one deliverable — and where any mismatch between the two (frame rate, or plain length) gets reconciled before the two are locked together. Common cases: a video exported at one frame rate needs to marry up with audio recorded or mixed separately; a translated or re-recorded audio track is longer or shorter than the picture it belongs to; footage shot at 24fps needs to play back at 25fps (or vice versa) for a different delivery spec, with the audio's pitch and pace following it.

Botverse Conform does this as a single MCP call. Point it at a video and an audio source, pick how frame-rate and duration differences should be resolved, and get back one fresh-encoded file.

Conform methods

Duration policy

Independently of the frame-rate method, the two sources might just be different lengths — a dubbed audio track that runs long, or a trimmed video that's shorter than its original audio. options.duration_policy decides what happens:

Splice: slates, bumpers, end cards, and joining clips

A second conform operation, alongside the video+audio mux above: conform_splice_from_url and conform_splice_media concatenate an ordered list of video and/or still-image segments into one continuous, fresh-encoded output. Where you put an image segment decides what it reads as — the mechanism underneath is identical either way:

Every segment is scaled to fit the output frame and letterboxed — not stretched — if its aspect ratio differs, so a portrait or square slate against widescreen footage doesn't distort. Any segment without its own audio (every still image, or an occasional silent video clip) gets a matching silent track synthesised automatically, so the output's audio never drops out mid-timeline. At least one segment must be a video — its resolution and frame rate become the output's defaults unless options.width/height/framerate override them. 2–8 segments per job.

Tools

Example — combine a picture and a replacement audio track

The simplest conform: video and audio are already at the same frame rate and the same length, and you just need them muxed into one deliverable.

// Agent calls:
{
  "name": "conform_from_url",
  "arguments": {
    "video_url": "https://cdn.example.com/footage/scene-04-picture.mp4",
    "audio_url": "https://cdn.example.com/footage/scene-04-remix.wav",
    "output_format": "mp4"
  }
}

// Botverse returns:
{
  "job_id": "job_9a1c2f4e",
  "status": "pending",
  "estimated_cost_usd": 0.30
}

Example — reconcile a frame-rate mismatch

Picture was delivered at 30fps; the target spec needs 24fps, and dropping every fifth frame would judder too much on this footage. Use interpolate_video for a motion-compensated retime:

{
  "name": "conform_from_url",
  "arguments": {
    "video_url": "https://cdn.example.com/footage/interview-30fps.mp4",
    "audio_url": "https://cdn.example.com/footage/interview-mixed.wav",
    "output_format": "mp4",
    "options": {
      "method": "interpolate_video",
      "target_framerate": 24
    }
  }
}

Example — true pulldown-style speed conform

Footage was shot at 24fps and needs to conform to a 25fps PAL-speed delivery: every frame is kept, but the whole thing plays back about 4% faster, and the audio needs to speed up by the same ratio to stay in sync.

{
  "name": "conform_from_url",
  "arguments": {
    "video_url": "https://cdn.example.com/footage/master-24fps.mov",
    "audio_url": "https://cdn.example.com/footage/master-mix-24fps.wav",
    "output_format": "mov_prores",
    "options": {
      "method": "speed_conform",
      "target_framerate": 25
    }
  }
}

// Botverse detects the source is 24fps and retimes both streams by 24/25 —
// output runs about 4% shorter in wall-clock time, same frame count, audio
// pitch/pace shifted to match.

Example — audio track runs long, keep the video's length

A dubbed or remixed audio track that's a few seconds longer than the picture. Use duration_policy: "match_video" so the video is never cut short and the extra audio is what gets truncated:

{
  "name": "conform_from_url",
  "arguments": {
    "video_url": "https://cdn.example.com/footage/promo-picture.mp4",
    "audio_url": "https://cdn.example.com/footage/promo-audio-fr.wav",
    "output_format": "mp4",
    "options": {
      "duration_policy": "match_video"
    }
  }
}

Poll for completion, then download

let status;
do {
  await new Promise(r => setTimeout(r, 5000));
  status = await callTool("get_job_status", { job_id: "job_9a1c2f4e" });
} while (!["complete", "failed"].includes(status.status));

if (status.status === "complete") {
  const { download_url } = await callTool("get_download_url", { job_id: "job_9a1c2f4e" });
}

From the command line

Same pipeline, no agent required:

botverse conform video.mp4 audio.wav --to mp4 \
  --method speed_conform --target-framerate 25

Example — a title card in front of the footage (head slate / bumper)

Put the image segment first. It holds for duration_seconds, then the video plays — with silent audio synthesised under the still so the track never drops out at the cut:

{
  "name": "conform_splice_from_url",
  "arguments": {
    "segments": [
      { "type": "image", "url": "https://cdn.example.com/title-card.png", "duration_seconds": 3 },
      { "type": "video", "url": "https://cdn.example.com/episode-04.mp4" }
    ],
    "output_format": "mp4"
  }
}

// Botverse returns:
{
  "job_id": "job_7f2b9a1c",
  "status": "pending",
  "segments": 2,
  "estimated_cost_usd": 0.30
}

Example — an end card after the footage (tail slate)

Same shape, image segment last instead of first:

{
  "name": "conform_splice_from_url",
  "arguments": {
    "segments": [
      { "type": "video", "url": "https://cdn.example.com/episode-04.mp4" },
      { "type": "image", "url": "https://cdn.example.com/contact-card.png", "duration_seconds": 5 }
    ],
    "output_format": "mp4"
  }
}

Example — joining two real clips (no image involved)

All-video segments splice the clips back-to-back — the same tool covers this with no image at all:

{
  "name": "conform_splice_media",
  "arguments": {
    "segments": [
      { "type": "video", "object_key": "inputs/…/part-a.mp4" },
      { "type": "video", "object_key": "inputs/…/part-b.mp4" }
    ],
    "output_format": "mp4"
  }
}

Which workflow to use