> ## Documentation Index > Fetch the complete documentation index at: https://docs.rigbox.dev/llms.txt > Use this file to discover all available pages before exploring further. # rig deploy > Deploy the project from its rig.yaml Available **local and workspace**. Reference source: CLI `0.13.0-rc.5` (`824aefd6ad16`). Deploys the project from the directory containing `rig.yaml`. On your laptop it targets `--workspace`, else the workspace bound in `.rig.lock`, else a new one. Inside a workspace it deploys into that workspace only: naming another workspace, `--bluegreen`, `--promote`, `--reimage`, and image deploys refuse before any work starts. ## Examples Deploy the project from the local directory containing rig.yaml. Review the deployment strategy before running. ```bash theme={null} rig deploy ``` Inspect the supported arguments before running this command. Help is offline and does not modify resources. ```bash theme={null} rig deploy --help ``` ## Local syntax ```text theme={null} Usage: rigbox deploy [OPTIONS] ``` ### Arguments and flags | Argument | Description | Details | | - | - | - | | `--from-dir` | Project directory containing rig.yaml. Defaults to cwd | Value: FROM\_DIR; Aliases: from | | `-w, --workspace` | Target workspace (name or id). When unset, reuses the single workspace bound in .rig.lock, or spawns a fresh one | Value: WORKSPACE | | `-n, --name` | Workspace name when spawning a new workspace | Value: NAME | | `--app` | Deploy only the named app (multi-app projects) | Value: APP | | `--watch` | Re-deploy on source changes (single-app projects only) | Default: false | | `--prune-stale` | Drop .rig.lock bindings to workspaces that no longer exist, then exit without deploying | Default: false | | `--stage` | Deployment stage. Also loads `.env.<stage>` and `.env.<stage>.local` (precedence: real env > .env.\.local > .env.local > .env.\ > .env). Falls back to \$RIG\_ENV | Value: STAGE | | `--no-env-file` | Skip auto-loading `.env` / `.env.<stage>` files | Default: false | | `--reimage` | Allow re-imaging a pre-existing workspace from a frozen image, discarding its current filesystem. Required to deploy a `reproducible: true` app into a workspace that wasn't created by this deploy and isn't already image-pinned | Default: false | | `--build-only` | Build/publish the frozen image(s) for `reproducible: true` apps and stop — no workspace, re-image, or rsync. A later `rig deploy` reuses the image | Default: false | | `--sync-only` | Sync every app's files but skip the restart pass — stage code without interrupting the running services | Default: false | | `--strategy` | Deploy managed app releases or explicitly build a workspace image | Value: STRATEGY; Choices: incremental, image | | `--deployment-id` | Stable deployment attempt ID for retries (incremental releases only) | Value: DEPLOYMENT\_ID | | `--bluegreen` | Stage a change before swinging production traffic at it; pair with `--promote` to swap atomically (zero-downtime). `<suffix>` stages an sibling app at `<base>-<suffix>` on its own subdomain. The special value `preview` instead clones the WHOLE workspace into a hidden preview (every app), overlays your new code, and serves it at `<app>-pv-<id>.rigbox.dev` — `--promote` then re-homes prod onto it. Pick by what the workspace IS: `preview` is for production SERVICE hosts (it swaps to a new workspace, so the old one is retired — open SSH sessions in it don't follow the swap). For a workspace people actively work IN (SSH, shells, dev processes), deploy in place or use `<suffix>` (same workspace, \~1s app restart) so sessions and workspace state survive. | Value: BLUEGREEN | | `--promote` | After staging via `--bluegreen <suffix>`, swap the production subdomain to the staged sibling in one atomic DB transaction (zero-downtime). The previous production app is retained as a rollback target until you delete it. Requires `--bluegreen`. | Default: false | | `--output` | Output format \[default: human, or `rig config set output …`] | Global; Value: OUTPUT; Choices: human, json, yaml, text, table | | `--query` | Filter output with a JMESPath expression | Global; Value: JMESPATH | | `--cli-input-json` | Read inputs from a JSON file or '-' (stdin) | Global; Value: FILE | | `--cli-input-yaml` | Read inputs from a YAML file or '-' (stdin) | Global; Value: FILE; Conflicts: cli\_input\_json | | `--generate-cli-skeleton` | Print a fillable input template, then exit | Global; Default: false | | `--no-pager` | Do not pipe help output through a pager | Global; Default: false | | `-v, --verbose` | Verbose diagnostics on stderr: per-phase deploy tracing, job-poll status, and transport commands. Useful for debugging a slow or stuck `rig deploy` | Global; Default: false | | `-h, --help` | Print help (see a summary with '-h') | | > **Complete command help** > > ```text theme={null} > Deploy the project from its rig.yaml > > Usage: rigbox deploy [OPTIONS] > > Options: > --from-dir > Project directory containing rig.yaml. Defaults to cwd > > [aliases: --from] > > --output > Output format [default: human, or `rig config set output …`] > > [possible values: human, json, yaml, text, table] > > --query > Filter output with a JMESPath expression > > -w, --workspace > Target workspace (name or id). When unset, reuses the single workspace bound in .rig.lock, or spawns a fresh one > > --cli-input-json > Read inputs from a JSON file or '-' (stdin) > > -n, --name > Workspace name when spawning a new workspace > > --app > Deploy only the named app (multi-app projects) > > --cli-input-yaml > Read inputs from a YAML file or '-' (stdin) > > --generate-cli-skeleton > Print a fillable input template, then exit > > --watch > Re-deploy on source changes (single-app projects only) > > --no-pager > Do not pipe help output through a pager > > --prune-stale > Drop .rig.lock bindings to workspaces that no longer exist, then exit without deploying > > --stage > Deployment stage. Also loads `.env.` and `.env..local` (precedence: real env > .env..local > .env.local > .env. > .env). Falls back to $RIG_ENV > > -v, --verbose > Verbose diagnostics on stderr: per-phase deploy tracing, job-poll status, and transport commands. Useful for debugging a slow or stuck `rig deploy` > > --no-env-file > Skip auto-loading `.env` / `.env.` files > > --reimage > Allow re-imaging a pre-existing workspace from a frozen image, discarding its current filesystem. Required to deploy a `reproducible: true` app into a workspace that wasn't created by this deploy and isn't already image-pinned > > --build-only > Build/publish the frozen image(s) for `reproducible: true` apps and stop — no workspace, re-image, or rsync. A later `rig deploy` reuses the image > > --sync-only > Sync every app's files but skip the restart pass — stage code without interrupting the running services > > --strategy > Deploy managed app releases or explicitly build a workspace image > > [possible values: incremental, image] > > --deployment-id > Stable deployment attempt ID for retries (incremental releases only) > > --bluegreen > Stage a change before swinging production traffic at it; pair with `--promote` to swap atomically (zero-downtime). `` stages an sibling app at `-` on its own subdomain. The special value `preview` instead clones the WHOLE workspace into a hidden preview (every app), overlays your new code, and serves it at `-pv-.rigbox.dev` — `--promote` then re-homes prod onto it. > > Pick by what the workspace IS: `preview` is for production SERVICE hosts (it swaps to a new workspace, so the old one is retired — open SSH sessions in it don't follow the swap). For a workspace people actively work IN (SSH, shells, dev processes), deploy in place or use `` (same workspace, ~1s app restart) so sessions and workspace state survive. > > --promote > After staging via `--bluegreen `, swap the production subdomain to the staged sibling in one atomic DB transaction (zero-downtime). The previous production app is retained as a rollback target until you delete it. > > Requires `--bluegreen`. > > -h, --help > Print help (see a summary with '-h') > ``` ## Workspace syntax ```text theme={null} Usage: rigbox deploy [OPTIONS] ``` ### Arguments and flags | Argument | Description | Details | | - | - | - | | `--from-dir` | Project directory containing rig.yaml. Defaults to cwd | Value: FROM\_DIR; Aliases: from | | `-w, --workspace` | Target workspace (name or id). When unset, reuses the single workspace bound in .rig.lock, or spawns a fresh one | Value: WORKSPACE | | `-n, --name` | Workspace name when spawning a new workspace | Value: NAME | | `--app` | Deploy only the named app (multi-app projects) | Value: APP | | `--watch` | Re-deploy on source changes (single-app projects only) | Default: false | | `--prune-stale` | Drop .rig.lock bindings to workspaces that no longer exist, then exit without deploying | Default: false | | `--stage` | Deployment stage. Also loads `.env.<stage>` and `.env.<stage>.local` (precedence: real env > .env.\.local > .env.local > .env.\ > .env). Falls back to \$RIG\_ENV | Value: STAGE | | `--no-env-file` | Skip auto-loading `.env` / `.env.<stage>` files | Default: false | | `--reimage` | Allow re-imaging a pre-existing workspace from a frozen image, discarding its current filesystem. Required to deploy a `reproducible: true` app into a workspace that wasn't created by this deploy and isn't already image-pinned | Default: false | | `--build-only` | Build/publish the frozen image(s) for `reproducible: true` apps and stop — no workspace, re-image, or rsync. A later `rig deploy` reuses the image | Default: false | | `--sync-only` | Sync every app's files but skip the restart pass — stage code without interrupting the running services | Default: false | | `--strategy` | Deploy managed app releases or explicitly build a workspace image | Value: STRATEGY; Choices: incremental, image | | `--deployment-id` | Stable deployment attempt ID for retries (incremental releases only) | Value: DEPLOYMENT\_ID | | `--bluegreen` | Stage a change before swinging production traffic at it; pair with `--promote` to swap atomically (zero-downtime). `<suffix>` stages an sibling app at `<base>-<suffix>` on its own subdomain. The special value `preview` instead clones the WHOLE workspace into a hidden preview (every app), overlays your new code, and serves it at `<app>-pv-<id>.rigbox.dev` — `--promote` then re-homes prod onto it. Pick by what the workspace IS: `preview` is for production SERVICE hosts (it swaps to a new workspace, so the old one is retired — open SSH sessions in it don't follow the swap). For a workspace people actively work IN (SSH, shells, dev processes), deploy in place or use `<suffix>` (same workspace, \~1s app restart) so sessions and workspace state survive. | Value: BLUEGREEN | | `--promote` | After staging via `--bluegreen <suffix>`, swap the production subdomain to the staged sibling in one atomic DB transaction (zero-downtime). The previous production app is retained as a rollback target until you delete it. Requires `--bluegreen`. | Default: false | | `--output` | Output format \[default: human, or `rig config set output …`] | Global; Value: OUTPUT; Choices: human, json, yaml, text, table | | `--query` | Filter output with a JMESPath expression | Global; Value: JMESPATH | | `--no-pager` | Do not pipe help output through a pager | Global; Default: false | | `-v, --verbose` | Verbose diagnostics on stderr (per-phase deploy tracing, job-poll status) | Global; Default: false | | `-h, --help` | Print help (see a summary with '-h') | | > **Complete command help** > > ```text theme={null} > Deploy the project from its rig.yaml > > Usage: rigbox deploy [OPTIONS] > > Options: > --from-dir > Project directory containing rig.yaml. Defaults to cwd > > [aliases: --from] > > --output > Output format [default: human, or `rig config set output …`] > > [possible values: human, json, yaml, text, table] > > --query > Filter output with a JMESPath expression > > -w, --workspace > Target workspace (name or id). When unset, reuses the single workspace bound in .rig.lock, or spawns a fresh one > > -n, --name > Workspace name when spawning a new workspace > > --no-pager > Do not pipe help output through a pager > > --app > Deploy only the named app (multi-app projects) > > -v, --verbose > Verbose diagnostics on stderr (per-phase deploy tracing, job-poll status) > > --watch > Re-deploy on source changes (single-app projects only) > > --prune-stale > Drop .rig.lock bindings to workspaces that no longer exist, then exit without deploying > > --stage > Deployment stage. Also loads `.env.` and `.env..local` (precedence: real env > .env..local > .env.local > .env. > .env). Falls back to $RIG_ENV > > --no-env-file > Skip auto-loading `.env` / `.env.` files > > --reimage > Allow re-imaging a pre-existing workspace from a frozen image, discarding its current filesystem. Required to deploy a `reproducible: true` app into a workspace that wasn't created by this deploy and isn't already image-pinned > > --build-only > Build/publish the frozen image(s) for `reproducible: true` apps and stop — no workspace, re-image, or rsync. A later `rig deploy` reuses the image > > --sync-only > Sync every app's files but skip the restart pass — stage code without interrupting the running services > > --strategy > Deploy managed app releases or explicitly build a workspace image > > [possible values: incremental, image] > > --deployment-id > Stable deployment attempt ID for retries (incremental releases only) > > --bluegreen > Stage a change before swinging production traffic at it; pair with `--promote` to swap atomically (zero-downtime). `` stages an sibling app at `-` on its own subdomain. The special value `preview` instead clones the WHOLE workspace into a hidden preview (every app), overlays your new code, and serves it at `-pv-.rigbox.dev` — `--promote` then re-homes prod onto it. > > Pick by what the workspace IS: `preview` is for production SERVICE hosts (it swaps to a new workspace, so the old one is retired — open SSH sessions in it don't follow the swap). For a workspace people actively work IN (SSH, shells, dev processes), deploy in place or use `` (same workspace, ~1s app restart) so sessions and workspace state survive. > > --promote > After staging via `--bluegreen `, swap the production subdomain to the staged sibling in one atomic DB transaction (zero-downtime). The previous production app is retained as a rollback target until you delete it. > > Requires `--bluegreen`. > > -h, --help > Print help (see a summary with '-h') > ``` This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.