Tools

The MCP server exposes 43 tools — a thin layer over the REST API. Each tool maps directly to one REST endpoint; the descriptions below double as prompt-engineering surface (it’s what the LLM in your client reads when deciding to call a tool).

The four exact-count tools (get_agent_search_count, get_personal_agent_count, get_agent_run_summary, get_space_agent_count) show their REST path directly — those endpoints are not yet published in the API reference. Use them for any “how many” question: the list-returning tools cap their results, so counting the rows they return under-reports on large accounts.

Agents

Tool	What it does	REST equivalent
`list_agents`	List all agents in the user’s organization.	List agents
`get_agent`	Get full detail on one agent.	Get an agent
`search_agents`	Search agents by name. Returns `{ agents, returned, limit, truncated }` — a full page means more may exist.	Search agents
`get_agent_search_count`	Exact number of agents matching a search term, server-side. Use instead of counting `search_agents` results.	`GET /api/v1/agent/search/count`
`get_personal_agent_count`	Exact number of agents in the caller’s personal space (agents with no space).	`GET /api/v1/agent/personal/count`

Runs

Tool	What it does	REST equivalent
`start_agent`	Start a run of an agent.	Start a run
`get_agent_runs`	List runs for an agent. Returns `{ runs, returned, limit, truncated }` — a full page means more may exist.	List runs for an agent
`get_agent_run_summary`	Exact run totals for an agent — overall count plus a per-status breakdown, across all run history.	`GET /api/v1/agent/{agentId}/runs/summary`
`get_run_status`	Status of a single run.	Get run status
`stop_agent`	Stop a run gracefully.	Stop a run
`kill_agent`	Force-kill a run.	Kill a run
`delete_run`	Delete a run from history.	Delete a run

Files

Tool	What it does	REST equivalent
`get_run_files`	List output files for a run.	List run output files
`get_file_download_url`	Signed URL to download one file.	Download a run file

Versions

Tool	What it does	REST equivalent
`get_agent_versions`	List all versions of an agent.	List versions
`restore_agent_version`	Roll back to a previous version.	Restore version

Schedules

Tool	What it does	REST equivalent
`list_agent_schedules`	List schedules for an agent.	List schedules
`get_agent_schedule`	Read one schedule.	Get schedule
`create_agent_schedule`	Create a CRON / RunEvery / RunOnce schedule.	Create schedule
`update_agent_schedule`	Modify a schedule.	Update schedule
`enable_agent_schedule`	Resume a paused schedule.	Enable schedule
`disable_agent_schedule`	Pause without deleting.	Disable schedule
`delete_agent_schedule`	Delete a schedule.	Delete schedule
`get_scheduled_runs`	Upcoming run instances from active schedules.	Upcoming schedules

Spaces

Tool	What it does	REST equivalent
`list_spaces`	All spaces visible to the user.	List spaces
`get_space`	Read one space.	Get a space
`get_space_agents`	Agents inside a space.	List space agents
`get_space_agent_count`	Exact number of agents in a space. Use instead of counting `get_space_agents` results.	`GET /api/v1/spaces/{spaceId}/agents/count`
`search_space_by_name`	Find a space by name.	Search spaces
`run_space_agents`	Run every agent in a space.	Run all

Billing

Tool	What it does	REST equivalent
`get_credits_balance`	Current credit balance.	Credits balance
`get_spending_summary`	Aggregate spend over a window.	Spending
`get_credit_history`	Transaction-level credit ledger.	Credit history
`get_agents_usage`	Per-agent usage roll-up.	Agents usage
`get_agent_cost_breakdown`	Per-agent cost detail.	Agent cost
`get_agent_runs_cost`	Cost per run for one agent.	Agent runs cost

Analytics

Tool	What it does	REST equivalent
`get_runs_summary`	Aggregate run counts by status.	Runs summary
`get_records_summary`	Aggregate records extracted.	Records summary
`get_run_diagnostics`	Failure diagnostics for one run.	Run diagnostics
`get_latest_failure`	Most recent failed run for an agent.	Latest failure

Agent Builder

Tool	What it does	REST equivalent
`start_agent_build`	Start an AI-powered build session from a natural-language prompt.	Start a session
`get_agent_build_status`	Poll a build session’s status (`processing`, `ready`, `completed`, `error`, `cancelled`).	Session status
`stop_agent_build`	Abort a build session early.	Stop session

Get started

Agent Builder

MCP

Agents

Control Center

Resources

Agents

Runs

Files

Versions

Schedules

Spaces

Billing

Analytics

Agent Builder

​Agents

​Runs

​Files

​Versions

​Schedules

​Spaces

​Billing

​Analytics

​Agent Builder

Agents

Runs

Files

Versions

Schedules

Spaces

Billing

Analytics

Agent Builder