deployments

Deploy your app to LlamaCloud and manage existing deployments.

Deployment names are the stable ids shown in the NAME column. Commands use the active profile’s project by default. Pass --project PROJECT on API-backed commands to override it for one command.

Usage

llamactl deployments [COMMAND] [options]

Commands:

• get [NAME]: List deployments, or show one deployment
• create: Create a deployment from an editor or YAML file
• edit [NAME]: Edit a deployment in your editor, or update from a YAML file
• apply -f FILE: Create or update a deployment from YAML
• template: Print a YAML scaffold for a new deployment
• delete NAME: Delete a deployment
• delete -f FILE: Delete the deployment named in a YAML file
• update NAME: Resolve a git ref again and start a new release
• history NAME: Show release history
• rollback NAME: Roll back to an earlier git SHA
• logs NAME: Fetch or stream deployment logs
• configure-git-remote NAME: Configure the push-mode git remote

Notes:

• -o json and -o yaml are for scripts. Status messages go to stderr; structured data goes to stdout.
• -f - reads YAML from stdin on commands that accept -f.
• repo_url: "" in apply YAML means push the current local working tree. Push-capable deployment mutations support --no-push when you have already pushed separately or want the server to use the revision it can already resolve.

Commands

Get

llamactl deployments get [NAME] [-o text|json|yaml|wide|template] [--project PROJECT]

With no NAME, lists deployments in the active project. With NAME, prints one deployment.

Output modes:

• text: Default table output
• wide: Table output with less common columns
• json: Machine-readable JSON
• yaml: Machine-readable YAML
• template: Apply-shaped YAML for one deployment only

Examples:

llamactl deployments get
llamactl deployments get invoice-agent -o yaml
llamactl deployments get invoice-agent -o template > deployment.yaml

Create

llamactl deployments create [-f FILE] [--no-push] [--project PROJECT]

Without -f, opens $EDITOR with a commented YAML scaffold. Save and close the file to create the deployment.

With -f FILE, creates from YAML without opening an editor. Use -f - to read from stdin.

Flags:

• -f, --filename FILE: YAML file, or - for stdin
• --no-push: Skip pushing local code for push-mode deployments
• --project PROJECT: Override the active project

Examples:

llamactl deployments create
llamactl deployments create -f deployment.yaml
llamactl deployments create -f - < deployment.yaml
llamactl deployments create --no-push

Edit

llamactl deployments edit [NAME] [-f FILE] [--no-push] [--project PROJECT]

Without -f, fetches the deployment, renders editable YAML, and opens $EDITOR. If NAME is omitted in a TTY, choose from existing deployments. Scripts should pass NAME.

With -f FILE, updates from YAML without opening an editor. If NAME is omitted, the YAML must include top-level name.

Flags:

• -f, --filename FILE: YAML file, or - for stdin
• --no-push: Skip pushing local code for push-mode deployments
• --project PROJECT: Override the active project

Examples:

llamactl deployments edit invoice-agent
llamactl deployments edit invoice-agent -f deployment.yaml
llamactl deployments edit -f deployment.yaml

Apply

llamactl deployments apply -f FILE [--dry-run] [--no-push] [--annotate-on-error] [--project PROJECT]

Applies deployment YAML declaratively. If the top-level name exists, apply updates that deployment. If it does not exist, apply creates it. YAML produced by deployments template or deployments get NAME -o template is ready for this command.

${VAR} references are resolved from the process environment at apply time. Masked secret values from read output are ignored so round-tripping a deployment does not overwrite existing secrets with placeholders.

Flags:

• -f, --filename FILE: Required YAML file, or - for stdin
• --dry-run: Validate and print the resolved payload without changing the deployment
• --no-push: Skip pushing local code for push-mode deployments
• --annotate-on-error: Write validation errors back into the YAML as comments
• --project PROJECT: Override the active project

Examples:

llamactl deployments template > deployment.yaml
llamactl deployments apply -f deployment.yaml
llamactl deployments apply -f deployment.yaml --dry-run
llamactl deployments apply -f deployment.yaml --annotate-on-error
llamactl deployments get invoice-agent -o template | llamactl deployments apply -f -

Template

llamactl deployments template

Prints a commented YAML scaffold for a new deployment. This command is offline and does not require auth. It uses local context when available, such as the current git remote, current branch, deployment config path, and secret names.

Example:

llamactl deployments template > deployment.yaml

Delete

llamactl deployments delete NAME [--project PROJECT]
llamactl deployments delete -f FILE [--project PROJECT]

Deletes a deployment immediately. Pass NAME directly, or pass -f FILE to read the deployment name from YAML.

Flags:

• -f, --filename FILE: YAML file containing top-level name
• --project PROJECT: Override the active project

Examples:

llamactl deployments delete invoice-agent
llamactl deployments delete -f deployment.yaml

Update

llamactl deployments update NAME [--git-ref REF] [--no-push] [--project PROJECT]

Resolves the deployment’s configured git ref again and starts a new release from the resulting commit. Use --git-ref to switch to a branch, tag, or commit before resolving.

For push-mode deployments, update mirrors local code before resolving the ref. Use --no-push if you already pushed separately or want to redeploy the revision already available to the server.

Flags:

• --git-ref REF: Branch, tag, or commit SHA to deploy
• --no-push: Skip mirroring local code for push-mode deployments
• --project PROJECT: Override the active project

Examples:

llamactl deployments update invoice-agent
llamactl deployments update invoice-agent --git-ref release-2026-05
llamactl deployments update invoice-agent --no-push

History

llamactl deployments history NAME [-o text|json|yaml|wide] [--project PROJECT]

Shows release history for a deployment, newest first.

Examples:

llamactl deployments history invoice-agent
llamactl deployments history invoice-agent -o yaml

Rollback

llamactl deployments rollback NAME [--git-sha SHA] [--project PROJECT]

Rolls a deployment back to a previous git SHA. In a TTY, omit --git-sha to choose from release history. Scripts should pass --git-sha.

Flags:

• --git-sha SHA: Git SHA to roll back to
• --project PROJECT: Override the active project

Examples:

llamactl deployments history invoice-agent
llamactl deployments rollback invoice-agent --git-sha 3f2a1c9
llamactl deployments rollback invoice-agent

Logs

llamactl deployments logs NAME [--follow] [--json] [--tail N] [--since-seconds N] [--include-init-containers] [--project PROJECT]

Fetches recent deployment logs and exits. Use --follow to keep streaming.

Flags:

• --follow, -f: Stream until interrupted
• --json: Emit one JSON log event per line
• --tail N: Number of log lines to fetch initially
• --since-seconds N: Only return logs newer than this many seconds
• --include-init-containers: Include init container logs
• --project PROJECT: Override the active project

Examples:

llamactl deployments logs invoice-agent
llamactl deployments logs invoice-agent --follow
llamactl deployments logs invoice-agent --tail 50 --since-seconds 3600
llamactl deployments logs invoice-agent --json

Configure Git Remote

llamactl deployments configure-git-remote NAME [--project PROJECT]

Configures an authenticated git remote for a push-mode deployment. The remote is named llamaagents-NAME.

Examples:

llamactl deployments configure-git-remote invoice-agent
git push llamaagents-invoice-agent

See also

• Getting started: Introduction
• Configure names, env, and UI: Deployment Config Reference
• Local dev server: llamactl serve