diff --git a/job_bundles/README.md b/job_bundles/README.md
index e360502..c21c4e2 100644
--- a/job_bundles/README.md
+++ b/job_bundles/README.md
@@ -109,6 +109,13 @@ S3 prefix, then distributes the hashing and data copies across a number of worke
 uses content-addressed storage for data files, users that later submit jobs with these files attached will not have to
 upload them.
 
+### FFmpeg movie from job output
+
+The [ffmpeg_movie_from_job_output](ffmpeg_movie_from_job_output) job bundle downloads the rendered output of another
+completed job in the same queue and uses FFmpeg to encode the image sequence into an MP4 video. This is useful as a
+post-processing utility — after a render job completes, submit this job with the source Job ID to automatically
+assemble the frames into a movie with configurable frame rate, quality, and resolution settings.
+
 ### SSH via SSM Managed Node
 
 The [ssh_to_smf](ssh_to_smf/README.md) job bundle registers a Deadline Cloud worker as an
diff --git a/job_bundles/ffmpeg_movie_from_job_output/.gitignore b/job_bundles/ffmpeg_movie_from_job_output/.gitignore
new file mode 100644
index 0000000..8ee0cf1
--- /dev/null
+++ b/job_bundles/ffmpeg_movie_from_job_output/.gitignore
@@ -0,0 +1 @@
+s3_settings.json
diff --git a/job_bundles/ffmpeg_movie_from_job_output/README.md b/job_bundles/ffmpeg_movie_from_job_output/README.md
new file mode 100644
index 0000000..bd27a2f
--- /dev/null
+++ b/job_bundles/ffmpeg_movie_from_job_output/README.md
@@ -0,0 +1,98 @@
+# FFmpeg Movie from Job Output
+
+## Introduction
+
+This job bundle downloads the rendered output of another completed job in the same queue
+and uses FFmpeg to encode the image sequence into an MP4 video file. It is useful as a
+post-processing utility — for example, after a Blender or Maya render job completes, you
+can submit this job to automatically assemble the frames into a movie.
+
+See also [ffmpeg_encode_video](../ffmpeg_encode_video) for a simpler sample that encodes
+a local image sequence without downloading from another job.
+
+## How it works
+
+A [pre-submission hook](https://github.com/aws-deadline/deadline-cloud/blob/mainline/docs/submission-hooks.md)
+(`inject_s3_settings.py`) runs at submission time on your workstation and looks up the
+queue's job attachment S3 bucket configuration. It writes the settings to a JSON file that
+gets uploaded as a job attachment, so the worker can access S3 without needing any
+Deadline Cloud API permissions.
+
+On the worker, the job installs the `deadline` Python library via pip in a job environment,
+then runs a single step that:
+
+1. Uses the `deadline.job_attachments` Python API to download the output files from the
+   source job's job attachments in S3.
+2. Auto-detects the image format from the downloaded files, sorts them alphabetically, and
+   encodes them into an H.264 MP4 video using FFmpeg with BT.709 color space metadata.
+
+## Prerequisites
+
+### Software
+
+The job requires FFmpeg (from conda-forge) and the Deadline Cloud Python library (installed
+via pip at runtime). On service-managed fleets, set the conda queue environment channel to
+`conda-forge`. The job's `CondaPackages` parameter defaults to `ffmpeg`.
+
+### Submission hooks
+
+This job bundle uses a pre-submission hook to inject S3 settings. Enable bundle hooks
+before submitting (one-time setup):
+
+```bash
+deadline config set settings.allow_bundle_hooks true
+```
+
+The hook runs on your local machine at submission time using your existing AWS credentials.
+No additional IAM permissions are needed on the queue role.
+
+### Source job requirements
+
+- The source job must have completed and produced output files via job attachments.
+- Both jobs must be in the same queue (they share the same job attachments S3 bucket).
+
+## Parameters
+
+| Parameter | Description | Default |
+|-----------|-------------|---------|
+| Source Job ID | The Job ID of the completed source job | (required) |
+| Source Step ID | Restrict download to a specific step's output | (empty = all) |
+| Frame Rate | Video frame rate in fps | 24 |
+| Pixel Format | Output pixel format (`yuv420p` or `yuv444p`) | yuv420p |
+| Encoding Preset | FFmpeg speed/compression tradeoff | medium |
+| Constant Rate Factor | H.264 quality (0 = lossless, 51 = worst, 17-18 ≈ visually lossless) | 18 |
+| Output Resolution | Optional WIDTHxHEIGHT override (e.g. `1920x1080`) | (empty = source) |
+| Output Filename | Name of the output video file | output.mp4 |
+| Output Directory | Where to save the video | output |
+
+## Example submission
+
+```bash
+# Enable bundle hooks (one-time setup)
+deadline config set settings.allow_bundle_hooks true
+
+# Submit via GUI
+deadline bundle gui-submit ffmpeg_movie_from_job_output/
+
+# Submit via CLI
+deadline bundle submit ffmpeg_movie_from_job_output/ \
+    -p SourceJobId=job-0123456789abcdef0123456789abcdef \
+    -p FrameRate=30 \
+    -p OutputFilename=my_render.mp4
+
+# Download only a specific step's output
+deadline bundle submit ffmpeg_movie_from_job_output/ \
+    -p SourceJobId=job-0123456789abcdef0123456789abcdef \
+    -p SourceStepId=step-0123456789abcdef0123456789abcdef
+```
+
+## Typical workflow
+
+1. Submit a render job (e.g. Blender, Maya) to your queue.
+2. Wait for the render job to complete.
+3. Copy the Job ID from Deadline Cloud Monitor.
+4. Submit this job bundle with the source Job ID.
+5. Download the output video from Deadline Cloud Monitor.
+
+You can also automate this by scripting the submission after the render job completes
+using `deadline job wait` followed by `deadline bundle submit`.
diff --git a/job_bundles/ffmpeg_movie_from_job_output/hooks.yaml b/job_bundles/ffmpeg_movie_from_job_output/hooks.yaml
new file mode 100644
index 0000000..5cd61ad
--- /dev/null
+++ b/job_bundles/ffmpeg_movie_from_job_output/hooks.yaml
@@ -0,0 +1,5 @@
+version: "1.0"
+preSubmission:
+  - command: python3
+    args: [inject_s3_settings.py]
+    timeout: 30
diff --git a/job_bundles/ffmpeg_movie_from_job_output/inject_s3_settings.py b/job_bundles/ffmpeg_movie_from_job_output/inject_s3_settings.py
new file mode 100644
index 0000000..104a219
--- /dev/null
+++ b/job_bundles/ffmpeg_movie_from_job_output/inject_s3_settings.py
@@ -0,0 +1,38 @@
+"""Pre-submission hook that writes job attachment S3 settings to a file in the bundle."""
+import json
+import os
+import sys
+
+from deadline.client import api
+
+metadata = json.load(sys.stdin)
+farm_id = metadata["farmId"]
+queue_id = metadata["queueId"]
+bundle_dir = metadata["jobBundleDir"]
+
+print(f"Looking up job attachment settings for queue {queue_id}...", file=sys.stderr)
+deadline = api.get_boto3_client("deadline")
+queue = deadline.get_queue(farmId=farm_id, queueId=queue_id)
+ja = queue.get("jobAttachmentSettings", {})
+
+if not ja:
+    print("ERROR: Queue has no job attachment settings configured.", file=sys.stderr)
+    sys.exit(1)
+
+bucket = ja["s3BucketName"]
+prefix = ja["rootPrefix"]
+print(f"S3 bucket: {bucket}, prefix: {prefix}", file=sys.stderr)
+
+# Write settings file into the bundle so it gets uploaded as a job attachment
+settings_path = os.path.join(bundle_dir, "s3_settings.json")
+with open(settings_path, "w") as f:
+    json.dump({"s3BucketName": bucket, "rootPrefix": prefix}, f)
+
+# Output asset reference so the file gets uploaded
+print(json.dumps({
+    "attachments": {
+        "assetReferences": {
+            "inputFilenames": [settings_path]
+        }
+    }
+}))
diff --git a/job_bundles/ffmpeg_movie_from_job_output/template.yaml b/job_bundles/ffmpeg_movie_from_job_output/template.yaml
new file mode 100644
index 0000000..42d1f32
--- /dev/null
+++ b/job_bundles/ffmpeg_movie_from_job_output/template.yaml
@@ -0,0 +1,282 @@
+specificationVersion: jobtemplate-2023-09
+name: FFmpeg Movie from Job Output
+description: |
+  This job downloads the output of another completed job in the same queue and uses
+  FFmpeg to encode the image sequence into a video file.
+
+  It requires FFmpeg from conda-forge and the Deadline Cloud Python library (installed
+  via pip at runtime). On service-managed fleets, use "conda-forge" as the CondaChannels
+  parameter for a conda queue environment.
+
+parameterDefinitions:
+# Source Job
+- name: SourceJobId
+  type: STRING
+  userInterface:
+    control: LINE_EDIT
+    label: Source Job ID
+    groupLabel: Source Job
+  description: >
+    The Job ID of a completed job in the same queue whose output images will be
+    assembled into a movie. Example: job-0123456789abcdef0123456789abcdef
+- name: SourceStepId
+  type: STRING
+  userInterface:
+    control: LINE_EDIT
+    label: Source Step ID (optional)
+    groupLabel: Source Job
+  default: ''
+  description: >
+    Optionally restrict the download to a specific step ID from the source job.
+    Leave empty to download all outputs.
+
+# Movie Settings
+- name: FrameRate
+  type: INT
+  userInterface:
+    control: SPIN_BOX
+    label: Frame Rate (fps)
+    groupLabel: Movie Settings
+  default: 24
+  minValue: 1
+  maxValue: 120
+  description: The frame rate of the output video.
+- name: PixelFormat
+  type: STRING
+  userInterface:
+    control: DROPDOWN_LIST
+    label: Pixel Format
+    groupLabel: Movie Settings
+  default: yuv420p
+  allowedValues: [yuv420p, yuv444p]
+  description: >
+    The pixel format for the output video. yuv420p is widely compatible;
+    yuv444p preserves more color detail.
+- name: EncodingPreset
+  type: STRING
+  userInterface:
+    control: DROPDOWN_LIST
+    label: Encoding Preset
+    groupLabel: Movie Settings
+  default: medium
+  allowedValues: [ultrafast, superfast, veryfast, faster, fast, medium, slow, slower, veryslow]
+  description: Controls the encoding speed to compression ratio.
+- name: ConstantRateFactor
+  type: INT
+  userInterface:
+    control: SPIN_BOX
+    label: Constant Rate Factor (CRF)
+    groupLabel: Movie Settings
+  default: 18
+  minValue: 0
+  maxValue: 51
+  description: >
+    Quality setting for H.264 encoding. 0 is lossless, 51 is worst quality.
+    17-18 is nearly visually lossless.
+- name: Resolution
+  type: STRING
+  userInterface:
+    control: LINE_EDIT
+    label: Output Resolution (optional)
+    groupLabel: Movie Settings
+  default: ''
+  description: >
+    Optional output resolution as WIDTHxHEIGHT (e.g. 1920x1080). Leave empty
+    to use the source image resolution.
+- name: OutputFilename
+  type: STRING
+  userInterface:
+    control: LINE_EDIT
+    label: Output Filename
+    groupLabel: Movie Settings
+  default: output.mp4
+  description: The filename for the output video.
+
+# Output
+- name: OutputDir
+  type: PATH
+  objectType: DIRECTORY
+  dataFlow: OUT
+  userInterface:
+    control: CHOOSE_DIRECTORY
+    label: Output Directory
+    groupLabel: Output
+  default: output
+  description: The directory where the output video will be saved.
+
+# Software Environment
+- name: CondaPackages
+  type: STRING
+  userInterface:
+    control: HIDDEN
+  default: ffmpeg
+  description: >
+    Conda packages required by this job. Requires a conda queue environment.
+- name: CondaChannels
+  type: STRING
+  userInterface:
+    control: HIDDEN
+  default: conda-forge
+  description: >
+    Conda channels to get packages from. Requires a conda queue environment.
+
+# Injected by pre-submission hook (inject_s3_settings.py)
+- name: S3SettingsFile
+  type: PATH
+  objectType: FILE
+  dataFlow: IN
+  userInterface:
+    control: HIDDEN
+  default: s3_settings.json
+  description: >
+    JSON file with S3 bucket settings. Created by the pre-submission hook.
+
+jobEnvironments:
+- name: InstallDeadline
+  description: Installs the Deadline Cloud Python library via pip.
+  script:
+    actions:
+      onEnter:
+        command: bash
+        args: ['{{Env.File.Setup}}']
+    embeddedFiles:
+    - name: Setup
+      type: TEXT
+      runnable: true
+      data: |
+        #!/bin/env bash
+        set -euo pipefail
+        pip install deadline
+- name: UnbufferedOutput
+  variables:
+    PYTHONUNBUFFERED: "True"
+
+steps:
+- name: EncodeMovie
+  script:
+    actions:
+      onRun:
+        command: bash
+        args: ['{{Task.File.Run}}']
+    embeddedFiles:
+    - name: Download
+      filename: download_outputs.py
+      type: TEXT
+      data: |
+        """Download output files from a source job using the job attachments API."""
+        import json
+        import os
+        import sys
+
+        from deadline.job_attachments.download import OutputDownloader
+        from deadline.job_attachments.models import (
+            FileConflictResolution,
+            JobAttachmentS3Settings,
+        )
+
+        farm_id = os.environ["DEADLINE_FARM_ID"]
+        queue_id = os.environ["DEADLINE_QUEUE_ID"]
+        job_id = "{{Param.SourceJobId}}"
+        download_dir = sys.argv[1]
+
+        with open("{{Param.S3SettingsFile}}") as f:
+            s3_cfg = json.load(f)
+        s3_settings = JobAttachmentS3Settings(
+            s3BucketName=s3_cfg["s3BucketName"],
+            rootPrefix=s3_cfg["rootPrefix"],
+        )
+
+        step_id = "{{Param.SourceStepId}}" or None
+
+        downloader = OutputDownloader(
+            s3_settings=s3_settings,
+            farm_id=farm_id,
+            queue_id=queue_id,
+            job_id=job_id,
+            step_id=step_id,
+        )
+        output_roots = list(downloader.get_output_paths_by_root().keys())
+        for root in output_roots:
+            downloader.set_root_path(root, download_dir)
+        stats = downloader.download_job_output(
+            file_conflict_resolution=FileConflictResolution.OVERWRITE,
+        )
+        print(f"Downloaded {stats.processed_files} files ({stats.processed_bytes} bytes)")
+    - name: Run
+      type: TEXT
+      runnable: true
+      data: |
+        #!/bin/env bash
+        set -xeuo pipefail
+
+        DOWNLOAD_DIR="$(mktemp -d)"
+        mkdir -p '{{Param.OutputDir}}'
+
+        # --- Download outputs from source job ---
+        echo "Downloading outputs from source job {{Param.SourceJobId}}..."
+        python3 '{{Task.File.Download}}' "$DOWNLOAD_DIR"
+
+        # --- Find images ---
+        # Auto-detect the most common image extension
+        IMAGE_EXTS="png exr jpg jpeg tga tiff tif dpx hdr bmp"
+        BEST_EXT=""
+        BEST_COUNT=0
+        for ext in $IMAGE_EXTS; do
+          COUNT=$(find "$DOWNLOAD_DIR" -type f -name "*.$ext" | wc -l)
+          if [ "$COUNT" -gt "$BEST_COUNT" ]; then
+            BEST_COUNT=$COUNT
+            BEST_EXT=$ext
+          fi
+        done
+
+        if [ "$BEST_COUNT" -eq 0 ]; then
+          echo "ERROR: No image files found. Available files:"
+          find "$DOWNLOAD_DIR" -type f | head -20
+          exit 1
+        fi
+        echo "Detected $BEST_COUNT .$BEST_EXT files"
+
+        # --- Build sorted file list and encode ---
+        CONCAT_FILE="$(mktemp)"
+        find "$DOWNLOAD_DIR" -type f -name "*.$BEST_EXT" | sort | while IFS= read -r img; do
+          echo "file '$img'"
+        done > "$CONCAT_FILE"
+
+        echo "First images:"; head -3 "$CONCAT_FILE"
+        echo "Last images:"; tail -3 "$CONCAT_FILE"
+
+        # Build scale filter
+        SCALE_FILTER="scale=in_color_matrix=bt709:out_color_matrix=bt709"
+        RESOLUTION='{{Param.Resolution}}'
+        if [ -n "$RESOLUTION" ]; then
+          WIDTH="${RESOLUTION%%x*}"
+          HEIGHT="${RESOLUTION##*x}"
+          SCALE_FILTER="scale=${WIDTH}:${HEIGHT}:in_color_matrix=bt709:out_color_matrix=bt709"
+        fi
+
+        FFMPEG_ARGS=(
+          -y
+          -f concat -safe 0
+          -r {{Param.FrameRate}}
+          -i "$CONCAT_FILE"
+          -pix_fmt {{Param.PixelFormat}}
+          -vf "$SCALE_FILTER"
+          -c:v libx264
+          -preset {{Param.EncodingPreset}}
+          -crf {{Param.ConstantRateFactor}}
+          -color_range tv
+          -colorspace bt709
+          -color_primaries bt709
+          -color_trc iec61966-2-1
+          -movflags faststart
+          '{{Param.OutputDir}}/{{Param.OutputFilename}}'
+        )
+
+        echo "Encoding video..."
+        ffmpeg "${FFMPEG_ARGS[@]}"
+        echo "Video saved to {{Param.OutputDir}}/{{Param.OutputFilename}}"
+  hostRequirements:
+    attributes:
+    - name: attr.worker.os.family
+      anyOf:
+      - linux