Skip to main content
Glama

experiment_manage

Manage pipeline runs by listing, creating, starting, pausing, resuming, retrying, or killing experiments. Track costs, steps, and share results within FleetQ's AI agent platform.

Instructions

Manage experiments (pipeline runs). Actions: list, get (experiment_id), create (name, hypothesis, workflow_id), start (experiment_id), pause, resume, retry, retry_from_step (experiment_id, step_id), kill, valid_transitions (experiment_id), cost (experiment_id), steps (experiment_id), share (experiment_id).

Input Schema

TableJSON Schema
NameRequiredDescriptionDefault
actionYesAction to perform: list, get, create, start, pause, resume, retry, retry_from_step, kill, valid_transitions, cost, steps, share
statusNoFilter by status: draft, scoring, planning, building, executing, completed, killed, paused, etc.
limitNoMax results to return (default 10, max 100)
experiment_idYesThe experiment UUID
titleYesExperiment title
thesisNoExperiment thesis/hypothesis
trackNoExperiment track: growth, retention, revenue, engagement (default: growth)growth
budget_cap_creditsNoBudget cap in credits (default: 10000)
step_idYesThe playbook step UUID to retry from
reasonNoReason for killing the experiment
show_costsNoWhether to show cost data in the public view (for update action)
show_stagesNoWhether to show pipeline stages in the public view (for update action)
show_outputsNoWhether to show stage outputs in the public view (for update action)
expires_atNoISO8601 expiry datetime after which the share link is invalid. Pass null to remove expiry. (for update action)

Latest Blog Posts

MCP directory API

We provide all the information about MCP servers via our MCP API.

curl -X GET 'https://glama.ai/api/mcp/v1/servers/escapeboy/agent-fleet-o'

If you have feedback or need assistance with the MCP directory API, please join our Discord server