Tools

Tools are grouped into toolsets. All toolsets are on by default; narrow them to save context. Credits marks tools that start browser runs. Confirm marks tools that need confirm: true, which an agent should only pass when you asked for the change.

Every tool returns a short summary for the agent plus the full result as structured JSON. IDs come from earlier results, so an agent rarely needs to ask you for one.

core

The test loop: find or write a scenario, run it, read the evidence.

Tool	What it does	Notes
`get_credits`	Remaining credit balance
`list_projects`	Your projects, optionally filtered by name
`get_project`	One project's details and dashboard link
`list_scenarios`	Test scenarios, optionally for one project
`get_scenario`	One scenario: URL, prompt, specification and recent runs
`create_scenario`	Write a scenario from a plain-language prompt. Does not run it
`lint_scenario`	Check stored or draft instructions for problems that make runs fail or error
`improve_scenario`	AI rewrite of a scenario's instructions; `apply` saves it
`diagnose_run`	AI explanation of why a run failed or errored, with a scenario fix when the instructions caused it
`run_scenario`	Run a scenario in a real browser and wait for the verdict, bug count and report link	Credits
`run_project`	Run every scenario in a project or suite, optionally against a preview `baseUrl`	Credits
`wait_for_run`	Keep waiting for a run that is still going
`get_run`	Verdict, bug reports with reproduction steps, failed requests, console errors and screenshot names
`list_runs`	Recent runs, newest first
`get_session_events`	Recorded network, console, interaction, navigation and screenshot events, filtered and paged
`get_screenshot`	One screenshot from a run, as an image

run_scenario and run_project take an optional requestId. Repeating a request ID returns the first launch instead of charging again, and requireVerified also checks the server's proof that every required outcome passed.

manage

Tool	What it does	Notes
`create_project`	Create a project; its URL becomes the production environment
`update_project`	Rename a project or change its URL or description
`delete_project`	Delete a project with its scenarios and run history	Confirm
`list_environments`	A project's environments; pass a name as `environment` to run tools
`update_scenario`	Change a scenario's name, location, prompt or settings
`delete_scenario`	Delete a scenario	Confirm
`list_suites`	A project's named suites, for `run_project`
`suggest_suites`	AI proposal for grouping a project's scenarios into suites. Read-only
`get_run_logs`	Execution logs, for runs that ended in `error`
`rerun`	Run a previous run again with its original or current configuration	Credits
`repeat_scenario`	Run one scenario several times to check flakiness; `dryRun` shows the quote first	Credits
`get_scenario_repeat`	Read or keep waiting for a repeat request
`list_sessions`	Recorded browser sessions
`get_session`	A session's bug reports and event counts

discovery

Tool	What it does	Notes
`start_discovery`	Let the agent explore the app and draft scenarios for what it finds	Credits
`get_discovery`	Read a discovery run and its drafts, optionally waiting for it
`accept_discovery_drafts`	Turn reviewed drafts into scenarios; accepting twice never duplicates
`list_coverage`	What the project's tests cover, by area and role
`get_coverage`	One coverage item and its run history

defects

Tool	What it does	Notes
`list_defects`	Bugs Monito reported, grouped across runs
`get_defect`	A defect's occurrences, review history and fix verifications
`compare_defect_occurrence`	Compare one occurrence with a defect
`defect_feedback`	Mark a defect confirmed, false positive, known issue, fixed or unreviewed
`link_defect`	Attach an occurrence to a defect
`verify_defect_fix`	Rerun the defect's original scenario to check the fix	Credits
`get_reliability`	Pass rates and flaky scenarios over 7 or 30 days

automations

Tool	What it does	Notes
`list_automations`	Schedules, deploy webhooks and CI triggers
`create_automation`	Schedule a project's tests (hourly to weekly, UTC)	Credits
`update_automation`	Reschedule, retarget, enable or disable
`delete_automation`	Delete an automation	Confirm

fixtures

Tool	What it does	Notes
`list_fixtures`	Reusable test data setup and cleanup
`get_fixture`	A fixture's definition, current or at one revision
`create_fixture`	Create a fixture
`update_fixture`	Save a new fixture revision; a concurrent edit is reported, not overwritten
`delete_fixture`	Delete a fixture at an exact revision	Confirm
`list_credentials`	An environment's saved login roles. Names only, never secrets
`get_visual_baseline`	A visual reference and its approval history
`approve_visual_baseline`	Use an inspected image as the reference future runs compare against

ci (local only)

Available only on the local server, because they read the repository, run gh and write .github/workflows/monito.yml. They follow the CI guide and need monito auth login and gh auth login.

Tool	What it does	Notes
`ci_plan`	Inspect the repository and plan the GitHub Actions integration. Read-only
`ci_apply`	Write the workflow and register the trust binding. Does not commit or push	Confirm
`ci_verify`	Check the setup; `remoteAuth` proves sign-in without a run, `run` starts a real run	Credits and Confirm with `run`
`ci_cleanup`	Plan removal of the integration; with `apply`, remove it	Confirm with `apply`

Not available over MCP

Some CLI commands have no tool on purpose: signing in and configuration (MCP clients sign in with OAuth), monito auth create-key and API key management (they would put a secret in the agent's context), and monito credential set (login secrets never pass through tool arguments). Use the dashboard or the CLI for those.

On this page