Launchmatic CLI
v0.10.1Deploy, test, and manage everything from your terminal. Zero config, instant feedback.
Install
npm i -g @launchmatic/cli
Core
lm login
Authenticate with your Launchmatic accountlm init
Initialize a service in the current directory — use -n for a service name, -r for a subdirectory root (e.g. apps/api)lm deploy
Build and deploy the current servicelm deploy --build-cmd "..." --start-cmd "..."
Override the detected build/start commands — the overrides persist for future deployslm deploy --pre-deploy-cmd "npx prisma migrate deploy"
Release-phase command run before rollout (migrations); failure aborts the deploylm quicklaunch(lm ql)
Detect, configure, and deploy in one command — auto-detects env vars with Requirement Connectorslm status
Show service and deployment statuslm logs
Stream runtime logs — add --build for build logs (-d to target a deployment)lm doctor
Scan the current project for common deployment issues before you shiplm destroy
Tear down a serviceMonorepo
Multi-service repos: bind every app in a workspace once, then deploy them together.
lm monorepo init(lm mono init)
Detect workspace services and write launchmatic.json — use -y to accept all defaultslm monorepo list(lm mono ls)
List the services bound in launchmatic.jsonlm up
Deploy every service whose files changed since the last push — Railway-style multi-service deploylm up --service web --service api
Deploy only the named serviceslm up --all
Skip the changed-only filter and deploy everythingEnvironment & Domains
lm env set KEY=VALUE [KEY2=VALUE2 ...]
Set one or more environment variableslm env list
List all environment variables (values masked)lm domains add example.com
Add a custom domain — prints the DNS record; SSL provisions automatically once it resolvesTemplates
Deploy a pre-built template — including instant managed images — without writing code first.
lm template list(lm tpl ls)
Browse available templates — filter with -c (fullstack, backend, frontend, ai, data, devtools)lm template deploy <slug>
Deploy a template — set required env vars inline with -e KEY=value, skip prompts with -yAI
Build with natural language and operate with conversational ops.
lm agent
Launch an interactive REPL — multi-step natural-language ops over your services (deploys, env vars, logs, domains, databases). Streams over SSE.lm agent -p "why did my last deploy fail?"
One-shot — run a single prompt and exit. Add --json to pipe into jq.lm agent --thread-id <id>
Resume an existing agent conversation across CLI sessionslm agent --service <id> --environment production
Scope the session to a specific service or environmentlm lightspeed(lm c)
Modify your service with AI — describe changes, AI generates code and deployslm c "add dark mode"
Quick inline modificationlm image generate "a hero illustration"(lm img gen)
Generate images with AI — -c for count (1–4), -a for aspect ratio, -o for output dirlm image usage
Show your image generation quotalm image history
Recent image generations (-l to limit)API Keys & Tokens
Issue and revoke programmatic credentials. For OAuth-based AI client connections, use the Connected Apps dashboard instead.
lm api-key list(lm api-key ls)
List your API keys with last-used timestampslm api-key create <name>
Create a new API key — full token is shown ONCE, store it immediatelylm api-key delete <keyId>(lm api-key rm)
Revoke an API keyDeployments & Previews
Per-service deploy history, rollbacks, and per-branch preview environments.
lm deployments list(lm deploys ls)
List recent deployments for the current servicelm deployments info <deploymentId>
Show deployment details — status, build logs, commit, durationlm deployments cancel <deploymentId>
Cancel a running deploymentlm deployments rollback <deploymentId>
Roll back to a previous deployment in one commandlm preview list(lm preview ls)
List active preview environmentslm preview create <branch>
Spin up a preview environment for a branchlm preview delete <previewId>(lm preview rm)
Tear down a preview environmentInfra & Usage
Inspect platform pods and aggregate resource usage. Most users don't need these — they're here when you do.
lm usage summary
Aggregated CPU, memory, network, and request stats for the current servicelm usage history
Time-series resource snapshots for trend analysislm infra list(lm infra ls)
List pods for a team (admin / debugging)lm infra logs <namespace> <pod>
Get logs straight from a specific podlm infra describe <namespace> <pod>
Detailed pod information — events, conditions, volumeslm infra restart <namespace> <pod>
Restart (delete and reschedule) a podlm infra delete <namespace> <pod>(lm infra rm)
Delete a podBrowser Automation
Powered by Playwright. Chromium auto-downloads on first use — zero setup required.
lm browser screenshot [url](lm b ss)
Capture a screenshot — supports device emulation, full-page, and delaylm b screenshot --device 'iPhone 14'
Screenshot with mobile device emulationlm b screenshot --full -o hero.png
Full-page screenshot saved to hero.pnglm browser pdf [url]
Generate a PDF — configurable format and orientationlm browser open [url]
Open a headed browser for interactive inspectionlm browser test [url]
Run smoke tests — load time, errors, HTTPS, contentlm b test --check-links --check-a11y
Extended tests with link validation and accessibility checkslm browser wait [url]
Poll until a page is reachable — great after deploylm browser codegen [url]
Record browser actions and generate Playwright test codelm browser devices
List available device emulation profileslm b devices --filter iphone
Filter devices by nameWorkflows
List and trigger Stirrup workflows running on your deployed services. The same workflows are also available to MCP clients via the agent's run_stirrup_workflow tool.
lm workflows list(lm wf ls)
List Stirrup workflows across every Stirrup service in your accountlm workflows run <serviceId>/<workflowId>
Trigger a workflow. Use --input '{...}' to pass a JSON payloadlm workflows runs <serviceId>
Show recent run history for a service's workflowsDatabase
Manage databases directly from the CLI — interactive shell, queries, dumps, and more.
lm db list(lm db ls)
List all databases with status and linked serviceslm db info <dbId>
Show database details and connection infolm db connect <dbId>(lm db url)
Get the full connection URL (with credentials)lm db credentials <dbId>(lm db creds)
Show host, port, username, password, and internal DNSlm db shell <dbId>(lm db sh)
Interactive SQL shell with psql-style meta-commands (\dt, \d, \l)lm db query <dbId> "SELECT * FROM users"(lm db q)
Execute a SQL query and print resultslm db tables <dbId>
List tables with sizes and row countslm db create my-db
Create a new PostgreSQL database — add --service <serviceId> to inject its URL into a service on creationlm db create my-cache -e redis
Create a Redis database (-e: postgresql, postgis, redis, mongodb)lm db link <dbId> <serviceId>
Link database to service (injects connection URL as env var)lm db dump <dbId>
Dump database to stdout or file (pg_dump)lm db seed <dbId> schema.sql
Run a SQL file against the databaselm db delete <dbId>(lm db rm)
Delete a databaseGitHub Repository
Manage your connected GitHub repo directly from the CLI. Auto-detects repo from git remote.
lm repo view
View repo info — stars, language, descriptionlm repo view --web
Open repository in browserlm repo prs
List pull requests (filterable by state)lm repo issues
List issues (filterable by state and labels)lm repo issue-create -t 'Bug: ...'(lm repo ic)
Create a new issue from the command linelm repo branches(lm repo br)
List branches, highlight currentlm repo commits(lm repo log)
Recent commits with author and datelm repo releases(lm repo rl)
List releases with tags and noteslm repo actions(lm repo runs)
View recent GitHub Actions workflow runslm repo browse [path](lm repo web)
Open repo or specific file in browserlm repo clone owner/repo
Shallow clone a GitHub repositorylm repo diff
Show uncommitted changes (git diff --stat)lm repo stash [push|pop|list]
Manage git stashPrefer your editor over the CLI?
The same agent and 11 tools that power lm agent are also reachable from Claude Code, Cursor, Codex, and any MCP client over OAuth at mcp.launchmatic.io. See the MCP setup docs.
Browser automation — zero config
Launchmatic auto-downloads Chromium on first use of lm browser — no manual setup needed. If you already have Chrome or Chromium installed, it uses that instead.
All browser commands auto-detect the URL from your .launchmatic.json context if not provided.