Render CLI Reference

Open the Render docs in your browser

Usage:

Examples:

Options:

Global options only

List environments for a specified project in the active workspace. In interactive mode you can view each environment's individual services.

Usage:

Examples:

Options:

Global options only

Open a redis-cli or valkey-cli session for a Render Key Value instance. This command only supports interactive mode.

You can optionally pass the key value ID or name as an argument. To pass arguments to redis-cli or valkey-cli, use: render kv-cli [keyValueID|keyValueName] -- [redis-cli args]

Usage:

Examples:

Options:

Global options only

Log in to Render using the Render Dashboard

Usage:

Examples:

Options:

Global options only

Log out of Render

Usage:

Options:

Global options only

View logs for services and datastores.

Use flags to filter logs by resource, instance, time, text, level, type, host, status code, method, or path. Unlike in the Render Dashboard, you can view logs for multiple resources at once.

In interactive mode you can update the filters and view logs in real time, or set --tail=true to stream new logs.

Usage:

Examples:

Options:

Global options, plus:

Open a pgcli session to a Render Postgres database instance. This command only supports interactive mode.

You can optionally pass a database ID or name as an argument. To pass arguments to pgcli, use: render pgcli [postgresID|postgresName] -- [pgcli args]

Usage:

Examples:

Options:

Global options only

Browse projects in the active workspace. In interactive mode, select a project to view its environments.

Usage:

Examples:

Options:

Global options only

Open a psql session to a Render Postgres database.

Optionally pass the database ID or name as an argument. To pass arguments to psql, use: render psql [postgresID|postgresName] -- [psql args]

For non-interactive usage, use the --command flag: render psql [postgresID|postgresName] -c "SELECT * FROM users;" -o text

Additional psql flags can be passed after --: render psql [postgresID|postgresName] -c "SELECT 1;" -o json -- --csv -q

Usage:

Examples:

Options:

Global options, plus:

Restart a service by resource ID

Usage:

Examples:

Options:

Global options only

SSH into a service instance.

Specify a service ID, service name, or instance ID. Put SSH arguments after --: render ssh [serviceID|serviceName|instanceID] -- [ssh args]

For non-interactive use, specify a service or instance and set --output text. With a service, SSH connects to any available instance. The command streams SSH output without formatting. The json and yaml output formats stream the same raw output as text.

SSH uses your local SSH keys. Add the public key to your Render account before you connect. See https://render.com/docs/ssh#ssh-setup.

Usage:

Examples:

Options:

Global options, plus:

Display information about the current user

Usage:

Examples:

Options:

Global options only

List workspaces available to your account

Usage:

Examples:

Options:

Global options only

Validate a Blueprint file for errors before committing.

Validates:

• YAML syntax
• Schema validation (Required fields, types)
• Semantic validation (valid plans, regions, etc.)
• Conflict checking against existing resources

Usage:

Examples:

Options:

Global options, plus:

Cancel a running deploy

Usage:

Examples:

Options:

Global options only

Trigger a service deploy and stream logs in real time

Usage:

Examples:

Options:

Global options, plus:

List deploys for a service

Usage:

Examples:

Options:

Global options only

Cancel a running job

Usage:

Examples:

Options:

Global options only

Create a new job for a service

Usage:

Examples:

Options:

Global options, plus:

List jobs for a service

Usage:

Examples:

Options:

Global options only

Create a new Render Key Value instance.

In interactive mode, a prompt guides you through each option one at a time. In non-interactive mode (--output text/json/yaml), flags use defaults if not supplied. Use --confirm to skip all prompts (including final confirmation) and create immediately. Output will be human-readable; use --output json/yaml/text for machine-readable output.

Usage:

Examples:

Options:

Global options, plus:

Delete a Render Key Value instance.

Without --confirm, this command previews what would be deleted and makes no changes. Pass --confirm to actually delete the instance.

The positional argument accepts either a Key Value ID (red-...) or a name. If the name matches more than one instance, narrow the search with --environment <id|name>, or pass the Key Value ID directly.

Name lookup is scoped to your active workspace. If a name isn't found, switch workspaces with 'render workspace set <name|ID>' and try again, or pass the Key Value ID instead (which works across workspaces).

Usage:

Examples:

Options:

Global options, plus:

Get details and connection info for a Render Key Value instance.

The positional argument accepts either a Key Value ID (red-...) or a name. If the name matches more than one instance, narrow the search with --project <id|name>, --environment <id|name>, or pass the Key Value ID directly.

Name lookup is scoped to your active workspace. If a name isn't found, switch workspaces with 'render workspace set <name|ID>' and try again, or pass the Key Value ID instead (which works across workspaces).

Usage:

Examples:

Options:

Global options, plus:

List Render Key Value instances in the active workspace.

Use --project to narrow results to a single project, --environment to narrow to a single environment, or both — when both are supplied, the environment is resolved within that project.

Usage:

Examples:

Options:

Global options, plus:

Resume a suspended Render Key Value instance.

The positional argument accepts either a Key Value ID (red-...) or a name. If the name matches more than one instance, narrow the search with --environment <id|name>, or pass the Key Value ID directly.

Name lookup is scoped to your active workspace. If a name isn't found, switch workspaces with 'render workspace set <name|ID>' and try again, or pass the Key Value ID instead (which works across workspaces).

Usage:

Examples:

Options:

Global options, plus:

Suspend a Render Key Value instance.

Without --confirm, this command previews what would be suspended and makes no changes. Pass --confirm to actually suspend the instance.

The positional argument accepts either a Key Value ID (red-...) or a name. If the name matches more than one instance, narrow the search with --environment <id|name>, or pass the Key Value ID directly.

Name lookup is scoped to your active workspace. If a name isn't found, switch workspaces with 'render workspace set <name|ID>' and try again, or pass the Key Value ID instead (which works across workspaces).

Usage:

Examples:

Options:

Global options, plus:

Update an existing Render Key Value instance.

The positional argument is the target Key Value (ID red-... or name). At least one mutating flag must be supplied. Use --name to rename the instance; the positional argument always identifies the target and is never the new name.

Environment, project, workspace, and region are immutable. A Key Value instance cannot be moved between them; the --environment flag is for name disambiguation only.

Name lookup is scoped to your active workspace. If a name isn't found, switch workspaces with 'render workspace set <name|ID>' and try again, or pass the Key Value ID instead (which works across workspaces). If a name matches more than one instance, narrow the search with --environment <id|name>.

The --ip-allow-list flag replaces the server-side list; pass it once per entry. To remove all allow-list entries, pass --clear-ip-allow-list. The two flags are mutually exclusive.

Usage:

Examples:

Options:

Global options, plus:

Create a new Render Postgres database.

In interactive mode, a wizard guides you through the core choices for the database. The wizard owns those prompted values. Flag-only settings, such as --disk-size-gb, --database-name, --database-user, --ip-allow-list, and --read-replica, are still included in the create request.

Use --confirm to skip the wizard and create immediately from flags and defaults. When --confirm is used with the default interactive output mode, output is printed as text. Use --output json, yaml, or text for non-interactive output.

Usage:

Examples:

Options:

Global options, plus:

Delete a Render Postgres database.

Without --confirm, this command previews what would be deleted and makes no changes. Pass --confirm to actually delete the database.

The positional argument accepts either a Postgres ID (dpg-...) or a name. If the name matches more than one database, narrow the search with --project <id|name>, --environment <id|name>, or pass the Postgres ID directly.

Name lookup is scoped to your active workspace. If a name isn't found, switch workspaces with 'render workspace set <name|ID>' and try again, or pass the Postgres ID instead (which works across workspaces).

Usage:

Examples:

Options:

Global options, plus:

Get details and connection info for a Render Postgres database.

The positional argument accepts either a Postgres ID (dpg-...) or a name. If the name matches more than one database, narrow the search with --project <id|name>, --environment <id|name>, or pass the Postgres ID directly.

Name lookup is scoped to your active workspace. If a name isn't found, switch workspaces with 'render workspace set <name|ID>' and try again, or pass the Postgres ID instead (which works across workspaces).

Usage:

Examples:

Options:

Global options, plus:

List Render Postgres databases in the active workspace.

Use --project to narrow results to a single project, --environment to narrow to a single environment, or both — when both are supplied, the environment is resolved within that project.

Usage:

Aliases: ls

Examples:

Options:

Global options, plus:

Resume a suspended Render Postgres database.

The positional argument accepts either a Postgres ID (dpg-...) or a name. If the name matches more than one database, narrow the search with --project <id|name>, --environment <id|name>, or pass the Postgres ID directly.

Name lookup is scoped to your active workspace. If a name isn't found, switch workspaces with 'render workspace set <name|ID>' and try again, or pass the Postgres ID instead (which works across workspaces).

Usage:

Examples:

Options:

Global options, plus:

Suspend a Render Postgres database.

Without --confirm, this command previews what would be suspended and makes no changes. Pass --confirm to actually suspend the database.

The positional argument accepts either a Postgres ID (dpg-...) or a name. If the name matches more than one database, narrow the search with --project <id|name>, --environment <id|name>, or pass the Postgres ID directly.

Name lookup is scoped to your active workspace. If a name isn't found, switch workspaces with 'render workspace set <name|ID>' and try again, or pass the Postgres ID instead (which works across workspaces).

Usage:

Examples:

Options:

Global options, plus:

Update an existing Render Postgres database.

The positional argument is the target database (ID dpg-... or name). At least one mutating flag must be supplied. Use --name to rename the database; the positional argument always identifies the target and is never the new name.

Environment, project, workspace, and region are immutable. A database cannot be moved between them; the --project and --environment flags are for name disambiguation only.

Only the fields you pass are changed; everything else is left untouched.

Name lookup is scoped to your active workspace. If a name isn't found, switch workspaces with 'render workspace set <name|ID>' and try again, or pass the Postgres ID instead (which works across workspaces). If a name matches more than one database, narrow the search with --project <id|name> or --environment <id|name>.

The --ip-allow-list flag replaces the server-side list; pass it once per entry. To remove all allow-list entries, pass --clear-ip-allow-list. The two flags are mutually exclusive.

Usage:

Examples:

Options:

Global options, plus:

Manage sandboxes for your Render workspace.

Sandboxes are ephemeral compute environments for running code, agents, and experiments.

Every sandbox belongs to a sandbox group, which scopes it to a region. Manage groups with "render sandbox-groups".

Snapshot a sandbox with "render sandboxes snapshots create" and restore new sandboxes from it with "render sandboxes create --snapshot-id".

Usage:

Available commands:

Examples:

Options:

Global options only

Create a new sandbox in the current workspace.

Usage:

Examples:

Options:

Global options, plus:

List sandboxes in your workspace.

By default, terminated sandboxes are excluded. Use --all to include them, or --status to filter by specific statuses.

Usage:

Examples:

Options:

Global options, plus:

Terminate a running sandbox. This action is irreversible.

Without --confirm, this command previews what would be terminated and makes no changes. Pass --confirm to actually terminate the sandbox.

Usage:

Examples:

Options:

Global options only

Copy a file or directory between the local filesystem and a running sandbox.

Prefix the remote side with the sandbox ID and a colon, like scp. A relative sandbox path resolves inside the sandbox's home directory; an absolute path addresses the sandbox filesystem root. Directories are transferred as archives: uploading a local directory recreates it at the sandbox path, and downloading a sandbox directory recreates it under the local path. "cp" works as an alias.

Unlike cp, a directory destination is not nested into: copying ./src to sbx-abc123:/tmp/src puts src's contents at /tmp/src, not at /tmp/src/src. A single file copied to an existing local directory does land inside it.

Usage:

Aliases: cp

Examples:

Options:

Global options only

Run a single command in a running sandbox. Streams stdout and stderr as the command runs, then exits with the remote command's exit code.

Pass the command after a "--" separator so its own flags aren't parsed by the CLI.

Usage:

Examples:

Options:

Global options only

Manage snapshots of sandboxes in your Render workspace.

A snapshot captures a running sandbox so a new sandbox can be restored from it with "render sandboxes create --snapshot-id". A filesystem snapshot captures the writable filesystem and restores onto any plan. A runtime snapshot also captures memory and CPU state and restores only onto the plan of the source sandbox.

Snapshots belong to the sandbox group of their source sandbox. Getting, listing, or deleting a snapshot uses the active workspace's default group unless --group is provided.

Usage:

Available commands:

Examples:

Options:

Global options only

Capture a snapshot of a running sandbox.

The command returns as soon as the API accepts the request, while the snapshot status is still "creating". Pass --wait to return once it's "available", so a sandbox can start from it.

The sandbox keeps running after the capture.

Usage:

Examples:

Options:

Global options, plus:

Delete one sandbox snapshot by ID. This action is irreversible.

Uses the active workspace's default sandbox group unless --group is provided. Without --confirm, this command previews what would be deleted and makes no changes. A snapshot that is still being created cannot be deleted yet. An expired snapshot is reported as not found.

Usage:

Examples:

Options:

Global options, plus:

Get one sandbox snapshot by ID.

Uses the active workspace's default sandbox group unless --group is provided. An expired snapshot is reported as not found.

Usage:

Examples:

Options:

Global options, plus:

List the snapshots in a sandbox group, newest first.

Uses the active workspace's default sandbox group unless --group is provided. Deleted and expired snapshots are not listed.

Usage:

Examples:

Options:

Global options, plus:

Manage sandbox groups for your Render workspace.

Sandbox groups scope a workspace's sandboxes to a region and (optionally) an environment. The workspace's first group is its default until you set another with "render sandbox-groups default set".

Usage:

Available commands:

Examples:

Options:

Global options only

Create a sandbox group in your workspace.

Group names are unique within a workspace. The workspace's first group becomes its default, where "render sandboxes create" places sandboxes unless --group is given.

Usage:

Examples:

Options:

Global options, plus:

Manage the workspace's default sandbox group.

The default group is where "render sandboxes create" places sandboxes unless --group is given. A workspace has exactly one default group.

Usage:

Available commands:

Options:

Global options only

Make a sandbox group the default for its workspace.

The previous default stops being the default. Running this on the current default makes no changes. A suspended group cannot be made the default.

Usage:

Examples:

Options:

Global options only

Delete a sandbox group. This action is irreversible.

Deleting a group terminates its sandboxes and deletes its snapshots shortly after. Suspended groups cannot be deleted. The workspace's default group cannot be deleted; make another group the default with "render sandbox-groups default set <sandboxGroupId>" first. Without --confirm, show the group that would be deleted without making changes.

Usage:

Examples:

Options:

Global options only

List all sandbox groups in your workspace, newest first.

Usage:

Examples:

Options:

Global options only

Lists all services and datastores for the active workspace. In interactive mode, you can view logs, restart services, trigger deploys, SSH into instances, and connect to Render Postgres databases and Render Key Value instances.

Usage:

Aliases: service

Available commands:

Examples:

Options:

Global options, plus:

Create a new service on Render. This command only runs in non-interactive modes. Provide configuration options with flags.

Usage:

Examples:

Options:

Global options, plus:

Delete a service on Render.

Without --confirm, this command previews what would be deleted and makes no changes. Pass --confirm to actually delete the service.

The positional argument accepts a service ID (including srv- or crn- IDs) or a name. Name lookup is scoped to your active workspace. If the name matches more than one service, pass the service ID directly.

This command only runs non-interactively. If --output interactive is requested, it falls back to text output.

Usage:

Examples:

Options:

Global options only

List instances for a service

Usage:

Examples:

Options:

Global options only

Update a service on Render. This command only runs in non-interactive modes.

Provide configuration updates with flags.

Usage:

Examples:

Options:

Global options, plus:

Install and manage Render agent skills for AI coding tools such as Claude Code, Codex, OpenCode, and Cursor. Skills add deployment, debugging, and monitoring capabilities to your AI coding assistant.

Usage:

Available commands:

Options:

Global options only

Install Render agent skills from https://github.com/render-oss/skills to detected AI coding tools.

Supported tools: Claude Code, Codex, OpenCode, Cursor.

Skills can be installed at two scopes:

• user: Install to ~/.{tool}/skills/ (default, current user only)
• project: Install to ./.{tool}/skills/ (committed to Git, all collaborators)

By default an interactive prompt lets you pick scope, tools, and skills. Use --scope, --tool, and --skill flags to skip the prompts (useful for CI).

Usage:

Examples:

Options:

Global options, plus:

List installed Render skills and the AI tools they've been installed in. This reads from local state only, so the command doesn't require network access.

Use --scope to filter by installation scope (user or project).

Usage:

Examples:

Options:

Global options, plus:

Remove previously installed Render skills from detected AI coding tools.

By default an interactive prompt lets you pick which skills to remove. Use --skill and --all flags to skip the prompts.

Use --scope to remove from a specific scope (user or project).

Usage:

Examples:

Options:

Global options, plus:

Reinstall Render skills using the tool and skill selections saved by a previous "render skills install" run.

This fetches the latest version of each selected skill from the skills repository, compares with installed versions, and updates any that have changed.

Use --scope to update skills at a specific scope (user or project).

Usage:

Examples:

Options:

Global options, plus:

Create a new workflow service on Render.

In interactive mode, a form guides you through the required fields. In non-interactive mode, provide all required config with flags.

Environment variables (non-interactive only for now):

• Pass individual vars with --env-var KEY=VALUE (repeatable).

• Load from one or more .env files with --env-file PATH (repeatable). Every

  listed file must exist.

• Inline --env-var values override values from --env-file.

Usage:

Examples:

Options:

Global options, plus:

Start a workflow service in development mode for local testing. Required input: -- <command to start a workflow service>

This command runs your workflow service locally on port 8120, allowing you to list and run tasks without deploying to Render. Task runs and their logs are stored in memory, so you can query them after tasks complete.

The command will spawn a new subprocess with your specified command whenever it needs to run a task or list the defined tasks.

To interact with the local task server:

• Use the --local flag with other task commands (e.g., 'render workflows tasks list --local')

• Or set RENDER_USE_LOCAL_DEV=true when using the workflow client SDK

To use a different port:

• Specify --port when starting the dev server

• Then use --port with other task commands, or set RENDER_LOCAL_DEV_URL in the SDK

Environment variables:

• A .env file in the current directory is loaded automatically if present

• Use --env-file to load one or more specific files (later files override earlier ones)

• Loaded values override variables inherited from the parent shell

Usage:

Examples:

Options:

Global options, plus:

Scaffold a new Render Workflows project with example tasks.

Creates a working example project with task definitions, dependencies, and a README with instructions for local development and Client SDK integration.

In interactive mode you'll be prompted to select a language, template, output directory, and optional features. Use --confirm to skip all prompts and accept defaults, or pass individual flags to skip specific prompts.

With --confirm or non-interactive output (-o text/json/yaml), dependencies are installed and Git is initialized by default. Pass --install-deps=false or --git=false to opt out.

Usage:

Examples:

Options:

Global options, plus:

List workflow services in your workspace

Usage:

Examples:

Options:

Global options only

List and release workflow versions

Usage:

Available commands:

Examples:

Options:

Global options only

List all versions of a workflow service.

Each time you release a workflow service, Render creates a new workflow version. A version represents a specific snapshot of your workflow service code and its registered tasks at the time of release.

This command displays all versions for a workflow, showing:

• Version ID

• Creation timestamp

• Associated tasks

In interactive mode, you will be prompted to select a workflow if not provided.

Usage:

Examples:

Options:

Global options only

Release a new version of a workflow service.

This command triggers a new release of your workflow service on Render. With a new release, Render:

1. Pulls the latest code from your repository (or a specific commit)

2. Builds your workflow service

3. Registers all tasks it finds in the service

4. Creates a new workflow version

You can optionally specify a commit ID to release a specific version of your code.

In interactive mode, you will be prompted to:

• Select a workflow if not provided

• Confirm the release

Usage:

Examples:

Options:

Global options, plus:

Cancel an in-progress task run.

Use --local to cancel a task run in the local workflow development server.

Usage:

Examples:

Options:

Global options, plus:

Start a task with the provided input. In non-interactive mode, provide input with --input or --input-file.

You can specify the task by its workflow slug and task name (e.g., my-workflow/my-task), either as a positional argument or with --task.

Input Format: The input should be a JSON array where each element is an argument to the task. For example, if your task takes two arguments, provide: ["arg1", "arg2"]

You can provide input via:

• --input with inline JSON

• --input-file with a path to a JSON file

In interactive mode, you will be prompted to select the task and provide the input.

Usage:

Examples:

Options:

Global options, plus:

List tasks and manage their runs

Usage:

Available commands:

Examples:

Options:

Global options, plus:

List all tasks defined in a workflow version.

Tasks are user-defined functions registered with the Render Workflows SDK. Each time you release a workflow service, Render creates a new workflow version and registers all tasks it finds in that version.

In interactive mode, you will be prompted to select a workflow if not provided. Only tasks from the workflow's default (latest released) version are listed. Pass --all-versions to include every version.

Local Development: When using the --local flag, you don't need to provide a workflow version ID. Instead, the command connects to your local dev server (default port 8120) to list tasks from your running workflow service. Start the dev server first with: render workflows dev -- "<your command>"

Usage:

Examples:

Options:

Global options, plus:

Manage runs of workflow tasks.

A task run represents a single execution of a task with specific input parameters. Use these commands to start new runs, view task run history, inspect details, and cancel in-progress runs.

Usage:

Available commands:

Examples:

Options:

Global options, plus:

Cancel an in-progress task run.

Use --local to cancel a task run in the local workflow development server.

Usage:

Examples:

Options:

Global options, plus:

List all execution runs for a specific task.

A task run represents a single execution of a task with specific input parameters. This command shows the history of all runs for a given task.

You can specify the task by its workflow slug and task name (e.g., my-workflow/my-task), either as a positional argument or with --task.

In interactive mode, you will be prompted to select a task if not provided.

Usage:

Examples:

Options:

Global options, plus:

Display detailed information about a specific task run execution.

This command shows comprehensive information about a task run, including:

• Task run ID and status

• Input parameters provided

• Output or error result

• Start and completion timestamps

The task run ID is returned when you execute a task with: render workflows tasks runs start

In interactive mode, you will be prompted to select a task run if not provided.

Usage:

Examples:

Options:

Global options, plus:

Start a task with the provided input. In non-interactive mode, provide input with --input or --input-file.

You can specify the task by its workflow slug and task name (e.g., my-workflow/my-task), either as a positional argument or with --task.

Input Format: The input should be a JSON array where each element is an argument to the task. For example, if your task takes two arguments, provide: ["arg1", "arg2"]

You can provide input via:

• --input with inline JSON

• --input-file with a path to a JSON file

In interactive mode, you will be prompted to select the task and provide the input.

Usage:

Examples:

Options:

Global options, plus:

List flows and manage their runs

Usage:

Available commands:

Options:

Global options only

Show details for a flow

Usage:

Examples:

Options:

Global options only

List flows accessible to your active workspace.

Flows are filtered by the active workspace. Narrow further with --flow-slug, --workflow, or --workflow-version. The flag can be repeated.

By default only flows from each workflow's default (latest released) version are listed. Pass --all-versions to include every version.

In interactive mode, you will be prompted to select a workflow if no filter is provided. Select a flow to view its runs or start a new run, then view a run's logs or results.

Usage:

Aliases: ls

Examples:

Options:

Global options, plus:

Start, list, and inspect flow runs

Usage:

Available commands:

Options:

Global options only

Cancel a running flow run

Usage:

Examples:

Options:

Global options only

List flow runs accessible to your active workspace.

At least one of --flow-slug, --flow, --workflow, or --workflow-version is required. Narrow further with --status. Each flag is repeatable.

Usage:

Aliases: ls

Examples:

Options:

Global options, plus:

Display a flow run's status, input, and result.

If the run failed because a replay didn't match its history, the output names the step and what changed. If it failed on a limit, the output names the limit and the run's value.

Usage:

Examples:

Options:

Global options only

Start a flow run with the provided input.

The flow is identified by its slug in the format workflow-slug/flow-name. An optional version can be appended (workflow-slug/flow-name:version); without a version, the latest release is used. Pass the slug as a positional argument or with --flow.

Input is a JSON value: a JSON array for positional arguments, or a JSON object for named parameters. Provide it inline with --input or load it from a file with --input-file.

Usage:

Examples:

Options:

Global options, plus:

List and inspect flow run steps

Usage:

Available commands:

Options:

Global options only

List the steps a flow run has recorded

Usage:

Aliases: ls

Examples:

Options:

Global options, plus:

Show detailed information about a flow run step

Usage:

Examples:

Options:

Global options only

Show the currently selected workspace

Usage:

Examples:

Options:

Global options only

Set the CLI's active workspace. All CLI commands run against the active workspace.

The active workspace is saved in cli.yaml in the Render CLI config directory. Set the RENDER_CLI_CONFIG_DIR environment variable to override this directory. If unspecified, the config file is saved in $HOME/.render/cli.yaml.

Usage:

Options:

Global options only