CLI command reference

The squasher CLI calls the same operations as the REST API from a terminal. Scripts and coding agents use it for repeatable, auditable work. Run squasher <command> --help for every flag.

Get the CLI

The CLI has no public install package yet. There is no npm, Homebrew, or crates.io package and no download. squasher update cannot download a new binary; it prints the command that rebuilds the CLI from source. The installer at squasher.ai/install.sh installs agent skills and the MCP server, not the CLI. Today, use the MCP server or the REST API for the same tasks.

The rest of this page is for teams that already have the squasher binary.

squasher login                          # Sign in in the browser (OAuth)
squasher login --device-code            # Sign in with a code, when there is no local browser
squasher login --key "$SQUASHER_API_KEY" # Store an API key, for CI and scripts
squasher config                         # Show the configuration and the credentials backend
squasher logout                         # Remove the stored credentials

squasher login opens your browser and waits on a local callback. squasher login --device-code prints a URL and a code to approve on any device. See CLI sign-in and integrations for storage and profiles.

Set the project

Most project commands need a project ID. The CLI takes it from the first of these that is set:

The --project <id> flag.
The SQUASHER_PROJECT_ID environment variable.
The active project of the profile, set with squasher projects use <id>.

squasher projects list
squasher projects use "$SQUASHER_PROJECT_ID"
squasher projects setup "$SQUASHER_PROJECT_ID"

projects setup shows whether the project has an ingest key and an event from the last 24 hours. It never prints a key or an event payload.

Global options

Option	Environment variable	Effect
`--json`		Print machine-readable JSON where the command supports it.
`--profile <name>`	`SQUASHER_PROFILE`	Use a named local profile for this command.
`--api-url <url>`	`SQUASHER_API_URL`	Change the API endpoint. The default is `https://api.squasher.ai`.
`--app-url <url>`	`SQUASHER_APP_URL`	Change the app endpoint for browser sign-in. The default is `https://app.squasher.ai`.

Run common tasks

# Find out which query answers a question
squasher query-guide get --project "$SQUASHER_PROJECT_ID"

# Summarize logs and latency for one service
squasher observe summary --project "$SQUASHER_PROJECT_ID" --service-name api

# List unresolved issues, newest first
squasher errors list --project "$SQUASHER_PROJECT_ID" --status unresolved --sort last_seen --limit 10

# Search error logs from the last hour
squasher logs query "timeout" --level error --since 1h

# Follow new error logs for five minutes
squasher logs tail --level error --duration 5m

# List incidents that are under investigation
squasher incidents list --project "$SQUASHER_PROJECT_ID" --status investigating

# Compare LCP at the 75th percentile by page
squasher vitals breakdown --project "$SQUASHER_PROJECT_ID" --metric LCP --dimension page_path --percentile p75 --period 24h

Add --json to any of these for output that a script can parse.

Commands

Run squasher <command> --help to see the subcommands and flags of each command.

Command	Alias	What it does
`login`	`signin`	Sign in with OAuth, or store an API key with `--key`.
`logout`	`signout`	Remove stored credentials.
`config`		Show the configuration and the credentials backend.
`profiles`		Create, list, and switch local profiles.
`settings`		View and edit local CLI settings.
`projects`	`services`	List projects, show setup status, set the active project.
`integrations`		Connect providers, and check provider health and resources.
`data-sources`		Manage data sources and integration-backed streams.
`log-connectors`	`log-drains`	Create, verify, and rotate keys for log connectors.
`alert-connectors`	`alerts`	Manage inbound alert connectors that open incidents.
`connectors`	`connector`	Browse the connectors that pull vendor audit logs.
`connections`	`connection`	Plan, connect, verify, and disconnect connector connections.
`push-sources`	`push-source`	Create push sources for signed webhooks and rotate secrets.
`infrastructure`	`infra`	Plan, install, and verify host telemetry collectors.
`github`		List, link, and unlink GitHub repositories.
`update`	`upgrade`	Print the command that rebuilds the CLI from source.

Observe

Command	Alias	What it does
`query-guide`		Show which query surface answers a question.
`observe`		Summarize logs and latency metrics for a project or service.
`errors`	`issues`	List, count, inspect, triage, and update issues.
`events`		Search error events.
`logs`		Search, query, tail, and forward logs, and read log insights.
`traces`		Inspect traces, operations, and service maps.
`metrics`		Inspect OpenTelemetry metrics.
`query`		Run safe read-only project queries.
`vitals`	`web-vitals`	Read Web Vitals summaries, details, and breakdowns.
`replays`	`sessions`	Search, read, and inspect session replays.
`deployments`		List and record deployments.
`ask`		Ask Squasher a question about a project in plain language.
`agent`		Change the Squasher Agent’s settings, memory, and skills.
`threads`		Search and manage saved investigation threads.
`messages`		Send messages to an investigation thread.

Respond

Command	Alias	What it does
`incidents`		Open, change, and inspect incidents and their timelines, and manage severity levels.
`on-call`		Read schedules, see who is on call, and add overrides.
`monitors`	`uptime`	Inspect and manage uptime monitors.
`automations`		Manage automations and their runs.
`remote-access`	`ssh`	Preview and request typed operations on private hosts. See Agent remote access.
`providers`		Read GitHub, Linear, Railway, and Vercel data linked to a project.

Manage

Command	Alias	What it does
`dashboards`		Manage dashboards as code.
`status-pages`		Manage status pages, components, and monitor links.
`maintenance`		Schedule, change, and delete maintenance windows.
`notifications`	`notification`	Manage notification destinations, routes, deliveries.
`api-keys`	`keys`	List, create, change, and revoke project API keys.
`subscribers`		Manage status page subscribers.
`sourcemaps`		Upload, test, and manage source maps.
`grouping-rules`	`grouping`	Set which issue new error events join.
`ingest-rules`	`drop-rules`	Drop or sample telemetry before it is stored.
`billing`		View billing plans.
`feedback`		Send feedback about Squasher to the Squasher team.
`ai`		Read AI and auto-fix observability summaries.
`agent-observations`	`observations`	Query agent observations.
`agent-traces`		Inspect per-trace span, token, cost, and error totals.
`scores`	`score`	Query quality scores on traces and generations.

Upload and test source maps

squasher sourcemaps upload sends the source maps of a release. Give it .map files, bundle files, or a build directory.

squasher sourcemaps upload --release 1.4.0 --service web dist/

A .map file uploads under its file name.
A .js, .mjs, or .cjs file uploads the map that its //# sourceMappingURL= comment names. The map uploads as <bundle file name>.map, so that Squasher can find it from a stack frame. The command prints the map that it used for each bundle.
A directory uploads the map of each bundle in it, and each other .map file in it.

squasher sourcemaps symbolicate resolves minified stack frames with the uploaded maps. Use it to make sure that the maps work before an error occurs.

squasher sourcemaps symbolicate --release 1.4.0 https://example.com/assets/app.js:1:20481
squasher sourcemaps symbolicate --release 1.4.0 --stack-file stack.txt

Each frame is url:line:col. With --stack-file, the command reads the frames from a stack trace (at fn (url:line:col) lines). The output shows one line for each frame: the original function and location, or unchanged and the reason. Use --debug-id when the bundle has a debug ID, and --context-lines to get lines of original source with --json.

Set grouping rules

Grouping rules decide which issue a new error event joins. Squasher applies the first enabled rule that matches. Issues that already exist keep their events.

squasher grouping-rules list
squasher grouping-rules create --name "Checkout timeouts" --action fingerprint \
  --exception-type TimeoutError --tag http.route=/checkout --template "{{ error.type }}-{{ tags.http.route }}"
squasher grouping-rules create --name "Vendor errors" --action stack-frames --frame-path node_modules/ --frames 2
squasher grouping-rules update <rule-id> --disable
squasher grouping-rules delete <rule-id> --yes

Action	What it does
`fingerprint`	Groups the events by the `--template` text.
`stack-frames`	Groups the events by the top in-app stack frames (`--frames`).
`merge-into`	Adds the events to an existing issue (`--merge-into iss_42`).

The conditions are --exception-type, --message-regex, --service, --frame-path, and --tag key=value. An event must match all the conditions that a rule sets. Use --position to set the order. Before you save a rule, try it on recent events. preview writes nothing:

squasher grouping-rules preview --hours 24 --rules-file rules.json

The rules file is a JSON array of rules in the API shape. Without --rules-file, the preview uses the saved rules. The output shows how many events would join a different issue and the largest groups. See Issues and error grouping.

Set ingest rules

Ingest rules drop or sample logs, spans, and metric points before Squasher stores them. Dropped data is not billed. Squasher applies the first enabled rule that matches. A change applies within about 60 seconds.

squasher ingest-rules preview --name "Health checks" --signal traces --action drop --span-name /health
squasher ingest-rules create --name "Health checks" --signal traces --action drop --span-name /health
squasher ingest-rules create --name "Debug logs" --signal logs --action sample --sample-percent 10 \
  --severity debug --service worker
squasher ingest-rules list
squasher ingest-rules reorder <rule-id> <rule-id>
squasher ingest-rules update <rule-id> --disable
squasher ingest-rules delete <rule-id> --yes

The conditions are --service, --severity (logs, repeatable), --span-name (traces), --metric-name (metrics, a prefix when it ends with *), --attr key=value, and --attr-regex key~pattern. Error spans and error logs are kept unless you add --include-errors true. preview writes nothing: it shows the matched volume over 7 days and the monthly saving. See Ingest rules.

Send feedback to Squasher

squasher feedback sends a bug report, an idea, or a docs problem to the Squasher team. It does not go to your project.

squasher feedback "The logs page does not show the service filter." --kind bug
echo "The on-call page does not explain overrides." | squasher feedback --kind docs

With no message, the command reads the message from standard input. It prints Receipt: <id>, or the full response with --json.

Option	What it does
`--kind`	`bug`, `idea`, `docs`, or `other`. The default is `other`.
`--title`	A short summary. The default is the first line of the message.
`--page`	The URL or path that the feedback is about.
`--tool`	The CLI command, MCP operation, or API endpoint that it is about.
`--anonymous`	Send without your credentials, through the anonymous Feedback.now API.

When you are signed in, the command uses POST /v1/feedback, which records your organization and project. When you are not signed in, or with --anonymous, it uses the anonymous Feedback.now endpoints. If your key can only send telemetry (events:ingest), the command sends the report anonymously and prints a note. The protocol commands are under squasher feedback now: discovery, policy, observe, submit, attach, and receipt. See Feedback API.

Get started

Guides

CLI command reference

Get the CLI

Set the project

Global options

Run common tasks

Commands

Observe

Respond

Manage

Upload and test source maps

Set grouping rules

Set ingest rules

Send feedback to Squasher

Next steps

CLI sign-in and integrations

Connect the MCP server

​Get the CLI

​Sign in

​Set the project

​Global options

​Run common tasks

​Commands

​Sign-in and setup

​Observe

​Respond

​Manage

​Upload and test source maps

​Set grouping rules

​Set ingest rules

​Send feedback to Squasher

​Next steps

CLI sign-in and integrations

Connect the MCP server

Get the CLI

Sign in

Set the project

Global options

Run common tasks

Commands

Sign-in and setup

Observe

Respond

Manage

Upload and test source maps

Set grouping rules

Set ingest rules

Send feedback to Squasher

Next steps