# Deadline Cloud

> Submit and manage rendering jobs on AWS Deadline Cloud, the AWS managed render farm service, from Python, the command line, or third party software.

- **Type:** MCP server
- **Install:** `agentstack add mcp-aws-deadline-deadline-cloud`
- **Verified:** Yes — security-reviewed for prompt injection and unsafe behavior
- **Seller:** [aws-deadline](https://agentstack.voostack.com/s/aws-deadline)
- **Installs:** 0
- **Category:** [Cloud & Infrastructure](https://agentstack.voostack.com/c/cloud-infrastructure)
- **Latest version:** 0.1.0
- **License:** Apache-2.0
- **Upstream author:** [aws-deadline](https://github.com/aws-deadline)
- **Source:** https://github.com/aws-deadline/deadline-cloud
- **Website:** https://aws-deadline.github.io/deadline-cloud/

## Install

```sh
agentstack add mcp-aws-deadline-deadline-cloud
```

Requires the [AgentStack CLI](https://agentstack.voostack.com/docs/cli). Works with Claude Code, Cursor, and any MCP-compatible agent.

## About

# AWS Deadline Cloud Client

### [User guide](https://aws-deadline.github.io/) | [Service documentation](https://docs.aws.amazon.com/deadline-cloud/) | [Deadline Cloud on GitHub](https://github.com/aws-deadline/) 

[](https://pypi.python.org/pypi/deadline)
[](https://pypi.python.org/pypi/deadline)
[](https://github.com/aws-deadline/deadline/blob/mainline/LICENSE)

[](https://github.com/aws-deadline/deadline-cloud/actions/workflows/dcm_integration_tests.yml?query=branch%3Amainline)

AWS Deadline Cloud client is a multi-purpose python library and command line tool for interacting with and submitting
[Open Job Description (OpenJD)][openjd] jobs to [AWS Deadline Cloud][deadline-cloud].

To support building workflows on top of AWS Deadline Cloud, it implements its own user interaction, job creation, file upload/download, and other useful
helpers around the service's API. It can function as a pipeline tool, a standalone GUI application, or even be embedded within other applications' runtimes.

Notable features include:
* A command-line interface with subcommands for querying your AWS Deadline Cloud resources, and submitting jobs to your AWS Deadline Cloud Farm.
* A library of functions for creating a job submission UI within any content creation tool that supports Python 3.9+ based plugins and
  the Qt GUI framework.
* A Model Context Protocol (MCP) server for AI assistant integration, enabling natural language interaction with AWS Deadline Cloud resources.

[cas]: https://en.wikipedia.org/wiki/Content-addressable_storage
[deadline-cloud]: https://docs.aws.amazon.com/deadline-cloud/latest/userguide/what-is-deadline-cloud.html
[deadline-cloud-monitor]: https://docs.aws.amazon.com/deadline-cloud/latest/userguide/working-with-deadline-monitor.html
[deadline-cloud-samples]: https://github.com/aws-deadline/deadline-cloud-samples
[deadline-jobs]: https://docs.aws.amazon.com/deadline-cloud/latest/userguide/deadline-cloud-jobs.html
[job-attachments]: https://docs.aws.amazon.com/deadline-cloud/latest/developerguide/build-job-attachments.html
[shared-storage]: https://docs.aws.amazon.com/deadline-cloud/latest/userguide/storage-shared.html
[job-bundles]: https://docs.aws.amazon.com/deadline-cloud/latest/developerguide/build-job-bundle.html
[openjd]: https://github.com/OpenJobDescription/openjd-specifications/wiki

## Compatibility

This library requires:

1. Python 3.9 through 3.14; and
2. Linux, Windows, or macOS operating system.

## Versioning

This package's version follows [Semantic Versioning 2.0](https://semver.org/), but is still considered to be in its
initial development, thus backwards incompatible versions are denoted by minor version bumps. To help illustrate how
versions will increment during this initial development stage, they are described below:

1. The MAJOR version is currently 0, indicating initial development.
2. The MINOR version is currently incremented when backwards incompatible changes are introduced to the public API.
3. The PATCH version is currently incremented when bug fixes or backwards compatible changes are introduced to the public API.

## Contributing

We welcome all contributions. Please see [CONTRIBUTING.md](https://github.com/aws-deadline/deadline-cloud/blob/mainline/CONTRIBUTING.md)
for guidance on how to contribute. Please report issues such as bugs, inaccurate or confusing information, and so on,
by making feature requests in the [issue tracker](https://github.com/aws-deadline/deadline-cloud/issues). We encourage
code contributions in the form of [pull requests](https://github.com/aws-deadline/deadline-cloud/pulls).

## Getting Started

The AWS Deadline Cloud client can be installed by the standard python packaging mechanisms:
```sh
$ pip install deadline
```

or if you want the optional gui dependencies:
```sh
$ pip install "deadline[gui]"
```

if you want the optional mcp dependencies:
```sh
$ pip install "deadline[mcp]"
```

or if you sign in with an AWS Console sign-in profile:
```sh
$ pip install "deadline[console]"
```

## Usage

After installation it can then be used as a command line tool:
```sh
$ deadline farm list
- farmId: farm-1234567890abcdefg
  displayName: my-first-farm
```

or as a python library:
```python
from deadline.client import api
api.list_farms()
# {'farms': [{'farmId': 'farm-1234567890abcdefg', 'displayName': 'my-first-farm', ...},]}
```

### Output format

The commands that accept an `--output` option (`auth status`, `config show`, `bundle gui-submit`, `job download-output`, `job download-input`, `job wait`, and `job logs`) choose their default format based on whether stdout is an interactive terminal. When you run them in a terminal you get the human-readable (`verbose`) output; when the output is piped, redirected, or has no TTY (for example in CI or when invoked by an agent) the default switches to `json`. Pass `--output` explicitly to override the detection — for example `--output verbose` to force human-readable output in a script, or `--output json` to force JSON in a terminal.

The `deadlinew` command can be used from GUIs to avoid displaying a terminal window in the background when on Windows.
You can use the `--redirect-output` option to write the terminal output to a file.
```sh
$ deadlinew --redirect-output out.txt farm list
$ cat out.txt
- farmId: farm-1234567890abcdefg
  displayName: my-first-farm
```

An example usage is to create a shortcut called "Deadline Settings" on your desktop that runs `C:\path\to\deadlinew.exe config gui`.
Opening the shortcut will show the Deadline Settings dialog without a terminal window behind it.

## Job-related Files
For job-related files and data, AWS Deadline Cloud supports either transferring files to AWS using job attachments or reading files from network storage that is shared between both your local workstation and your farm.

### Job attachments

Job attachments enable you to transfer files between your workstations and AWS Deadline Cloud using Amazon S3 buckets as
[content-addressed storage][cas] in your AWS account. The use of a content-addressed storage means that a file will never need
to be uploaded again once it has been uploaded once.

See [job attachments][job-attachments] for a more in-depth look at how files are uploaded, stored, and retrieved.

### Shared storage and storage profiles
Jobs can reference files that are stored on shared network storage. The Deadline Client uses a storage profile to determine which paths on the workstation are part of the network storage and do not need to be transferred using job attachments.

To use an existing storage profile with the Deadline Client, you can configure your default storage profile via CLI:

```sh
deadline config set settings.storage_profile_id sp-10b2e48ad6ac4fc88595dfcbef6271f2
```

Or with the configuration GUI:
```sh
deadline config gui
```

## Job Bundles

A job bundle is one of the tools that you can use to define jobs for AWS Deadline Cloud. They group an [Open Job Description (OpenJD)][openjd] template with
additional information such as files and directories that your jobs use with job attachments. You can use this package's command-line interface and/or
its Python interface to use a job bundle to submit jobs for a queue to run. Please see the [Job Bundles][job-bundles]
section of the AWS Deadline Cloud Developer Guide for detailed information on job bundles.

At a minimum, a job bundle is a folder that contains an [OpenJD][openjd] template as a file named `template.json` or `template.yaml`. However, it can optionally include:
1. An `asset_references.yaml` file - lists file inputs and outputs.
2. A `parameter_values.yaml` file - contains the selected values for the job template's parameters.
3. A `hooks.yaml` file - defines pre/post-submission hooks (see [Submission Hooks](#submission-hooks)).
4. Any number of additional files required for the job.

For example job bundles, visit the [samples repository][deadline-cloud-samples].

To submit a job bundle, you can run
```sh
$ deadline bundle submit 
```

or if you have the optional GUI components installed, you can load up a job bundle for submission by running:
```sh
$ deadline bundle gui-submit --browse
```

On submission, a job bundle will be created in the job history directory (default: `~/.deadline/job_history`).

### Submission Hooks

You can run custom scripts during job submission by adding a `hooks.yaml` file to your job bundle:

```yaml
preSubmission:
  - command: python3
    args: [validate_assets.py]
    timeout: 30

postSubmission:
  - command: python3
    args: [notify_team.py]
```

**Pre-submission hooks** run before files are uploaded and can:
- Validate job configuration
- Discover and add additional input files
- Modify submission parameters

**Post-submission hooks** run after job creation for notifications and integrations.

Hooks receive job metadata via environment variables (`DEADLINE_JOB_NAME`, `DEADLINE_FARM_ID`, etc.) and JSON on stdin. See [docs/submission-hooks.md](docs/submission-hooks.md) for full documentation.

## Configuration

You can see the current configuration by running:
```sh
$ deadline config show
```
and change the settings by running the associated `get`, `set` and `clear` commands.

If you need to parse the settings as json, you can specify the output by running:
```sh
$ deadline config show --output json
```
Which will output:
```sh
{"settings.config_file_path": "~/.deadline/config", "deadline-cloud-monitor.path": "", "defaults.aws_profile_name": "(default)", "settings.job_history_dir": "~/.deadline/job_history/(default)", "defaults.farm_id": "", "settings.storage_profile_id": "", "defaults.queue_id": "", "defaults.job_id": "", "settings.auto_accept": "false", "settings.conflict_resolution": "NOT_SELECTED", "settings.log_level": "WARNING", "telemetry.opt_out": "false", "telemetry.identifier": "", "defaults.job_attachments_file_system": "COPIED", "settings.s3_max_pool_connections": "50", "settings.small_file_threshold_multiplier": "20", "settings.known_asset_paths": "", "settings.locale": "", "settings.force_s3_check": "false", "settings.allow_bundle_hooks": "false", "settings.allow_environment_hooks": "false", "settings.submitter_update_notification": "true", "settings.max_retries_per_task": "5", "settings.max_failed_tasks_count": "20"}
```

To see a list of settings that can be configured, run:
```sh
$ deadline config --help
```

Or you can manage settings by a graphical user-interface if you have the optional GUI dependencies:
```sh
$ deadline config gui
```

By default, configuration of AWS Deadline Cloud is provided at `~/.deadline/config`, however this can be overridden by the `DEADLINE_CONFIG_FILE_PATH` environment variable.

## Authentication

In addition to the standard AWS credential mechanisms (AWS Profiles, instance profiles, and environment variables), AWS Deadline Cloud monitor and AWS Console sign-in credentials are also supported.

To view the currently configured credentials authentication status, run:

```sh
$ deadline auth status
    Profile Name: (default)
          Source: HOST_PROVIDED
          Status: AUTHENTICATED
API Availability: True
```

If the currently selected AWS Profile is set-up to use [AWS Deadline Cloud monitor][deadline-cloud-monitor] credentials, you can authenticate by logging in:

```sh
$ deadline auth login
```

and removing them by logging out:
```sh
$ deadline auth logout
```

AWS Console sign-in profiles, reported by `deadline auth status` with a source of
`AWS_CONSOLE_LOGIN`, are supported as well. Starting a session requires an interactive
browser sign-in, so `deadline auth login` opens AWS Deadline Cloud monitor to perform it
and waits for the profile to authenticate — the same handoff used for monitor profiles.

This requires AWS Deadline Cloud monitor to be installed and to have created the profile.
If it isn't configured, `deadline auth login` explains how to sign in instead:

* AWS Deadline Cloud monitor, using its "Login with AWS Console" option, or
* `aws login --profile `, using AWS CLI v2.

Once you are signed in, credentials refresh automatically until the session expires,
with no further sign-in and without invoking any external tool. That refresh does
require the `console` extra above (`pip install "deadline[console]"`).

`deadline auth logout` deletes the profile's cached token, which ends the session on
this workstation.

## Job Monitoring and Logs

### Waiting for Job Completion

After submitting a job, you can wait for it to complete using the `wait` command:

```sh
# Wait for a job to complete with default settings
$ deadline job wait --job-id job-12345

# Customize the maximum polling interval (default is 120 seconds)
# The polling interval starts at 0.5 seconds and doubles until reaching this maximum
$ deadline job wait --job-id job-12345 --max-poll-interval 30

# Set a timeout (default is 0, meaning no timeout)
$ deadline job wait --job-id job-12345 --timeout 3600

# Get the result in JSON format
$ deadline job wait --job-id job-12345 --output json
```

The command blocks until the job reaches a terminal state (SUCCEEDED, FAILED, CANCELED, SUSPENDED, NOT_COMPATIBLE), then returns information about the job's status and any failed tasks. It uses exponential backoff for polling, starting at 0.5 seconds and doubling the interval after each check until it reaches the maximum polling interval.

**Exit Codes:**
- `0` - Job succeeded
- `1` - Timeout waiting for job completion
- `2` - Job failed or has failed tasks
- `3` - Job was canceled
- `4` - Job was archived
- `5` - Job is not compatible

### Retrieving Job Logs

You can monitor job status and retrieve logs using the CLI. The logs lines are returned starting from the most recent log event with timestamps in ISO 8601 format:

```sh
# Get logs for a specific session
$ deadline job logs --session-id session-12345

# Get logs for a job (automatically selects session: ongoing sessions preferred, then most recently started/ended)
$ deadline job logs --job-id job-12345

# Limit the number of log lines returned to the 50 most recent.
$ deadline job logs --session-id session-12345 --limit 50

# Filter logs by time range
$ deadline job logs --session-id session-12345 --start-time 2023-01-01T12:00:00Z --end-time 2023-01-01T13:00:00Z

# Get logs in JSON format
$ deadline job logs --session-id session-12345 --output json

# Get logs with timestamps in local timezone (default is UTC)
$ deadline job logs --session-id session-12345 --timestamp-format local

# Get logs with explicit UTC timestamps (default behavior)
$ deadline job logs --session-id session-12345 --timestamp-format utc

# Get logs with relative timestamps
$ deadline job logs --session-id session-12345 --timestamp-format relative

# Combine timestamp format option with JSON output
$ deadline job logs --session-id session-12345 --timestamp-format local --output json

# Paginate through logs
$ deadline job logs --session-id session-12345 --next-token next-token-value
```

**Timestamp Format**: All timestamps are displayed in ISO 8601 format with full microsecond precision and timezone information:
- UTC format: `2025-07-03T10:49:33.821306+00:00`
- Local format: `2025-07-03T03:49:33.821306-07:00` (example for PST)

**Timestamp Format Options**:
- `--timestamp-format utc` (default): Display timestamps in UTC with `+00:00` offset
- `--timestamp-format local`: Display timestamps converted to your local system timezone
- `--timestamp-format relative`: Display timestamps relative to the session or session action start time

When using a Deadline Cloud monitor profile, the `job logs` command will use the Queue role credentials to read logs. Otherwise, the chosen profile credentials are used for all API invocations. This allows you to access logs with the appropriate permissions based on your authentication method.

## AWS Credentials Integration

You can use the Deadline Cloud client to obtain temporary AWS credentials for a queue and use them with the AWS CLI or SDK. This enab

…

## Source & license

This open-source MCP server is cataloged on AgentStack and links to its original source — we do not rehost the code.

- **Author:** [aws-deadline](https://github.com/aws-deadline)
- **Source:** [aws-deadline/deadline-cloud](https://github.com/aws-deadline/deadline-cloud)
- **License:** Apache-2.0
- **Homepage:** https://aws-deadline.github.io/deadline-cloud/

Install and usage instructions live in the source repository linked above.

## Pricing

- **Free** — Free

## Security capabilities

Automated source analysis of v0.1.0 — what this tool can access:

- **Network access:** no
- **Filesystem access:** no
- **Shell / process execution:** no
- **Environment & secrets:** no
- **Dynamic code execution:** no

*"Yes" means the capability is present in the source — more access means more to trust, not that it is unsafe.*


## Versions

- **0.1.0** — security scan: passed — Imported from the upstream source.

## Links

- Listing page: https://agentstack.voostack.com/l/mcp-aws-deadline-deadline-cloud
- Seller: https://agentstack.voostack.com/s/aws-deadline
- Browse the marketplace: https://agentstack.voostack.com/browse

---
Listed on AgentStack — the marketplace for AI agent skills and MCP servers. Every listing is security-reviewed. Creators keep 70%.