study run

Study ID (alias or UUID) - positional alias for —study (optional)

Workspace ID

Study ID

Iteration to run (defaults to latest on the study)

Person IDs/aliases (comma-separated or repeatable) Default: [].

Randomly sample N people from the matching pool (max 20 per dispatch - split into multiple slices for larger cohorts)

Use every AI person matching the filters (workspace-wide if no filters set) (capped at 20 per dispatch - split into multiple slices for larger cohorts)

Substring match against person name

Substring match against person bio

Substring match against person occupation (repeatable) Default: [].

Filter by gender (female, male, nonbinary; repeatable, OR semantics) Default: [].

Filter by 2-letter country code (repeatable) Default: [].

Minimum age (inclusive)

Maximum age (inclusive)

Filter by visibility: workspace (your workspace), shared (community-published), platform (admin-curated). Old values private / public are accepted and mapped to workspace / platform.

Simulation config ID (required for media unless every person has one)

Max interactions per participant (interactive / media only). Precedence: flag > iteration’s stored value > CLI default (20).

Max conversation turns per participant (chat studies only)

Allow chat agent to end the conversation early when goals are met (chat studies only)

Language code (e.g. en, sv)

Wait for all simulations to reach a terminal state before returning

Wait timeout in seconds (default 300; only with —wait)

Per-POST timeout in seconds for the create-participants + dispatch calls (default 120). Bump if study run times out client-side after seeding participants but before dispatch - those participants exist server-side and are surfaced under seeded_but_not_dispatched_* in the error envelope so the agent can resume.

Skip confirmation prompt

Run simulation with local browser (Playwright) instead of remote

Show browser window (local mode only)

Slow down actions by ms (local mode only)

Open Chrome DevTools (local mode only)

Enable detailed debug logging to stderr and ~/.ish/local-sim.log

Run N participants in parallel (local mode only). Browser: default all. Native iOS/Android: pools N auto-provisioned devices - simulators (iOS) / headless emulators from your AVDs (Android) - default 1, capped at 5, auto-sized to host RAM (and AVD count).

Local target platform: ‘web’ (Playwright), ‘android’ (adb emulator), or ‘ios’ (simctl+idb simulator). Defaults to the iteration’s platform.

Native local mode: path to an .apk (android) / .app (ios) to install, or an installed package/bundle id to launch. The extension implies —platform.

Participant ID (alias or UUID; from ish study run --json.participant_aliases[]) (optional)

Study ID (poll all simulations for study)

Workspace ID; accepted for consistency (workspace is inferred from —study/—iteration/<participant_id>)

Participant ID (alias or UUID; from ish study run --json.participant_aliases[]) (optional)

Study ID (wait for all participants in the study)

Iteration ID (wait for participants in this iteration only)

Max seconds to wait (default 300)

Workspace ID; accepted for consistency (workspace is inferred from —study/—iteration/<participant_id>)

Participant ID (alias or UUID; from ish study run --json.participant_aliases[]) (required)

Workspace ID; accepted for consistency (workspace is inferred from <participant_id>)

Participant to extend (alias or UUID). Must be in a terminal state (completed/failed/cancelled). (required)

Extra interactions past the source’s original cap (1-50; backend caps server-side) Default: “10”.

User message to inject as the new participant resumes. Accepts inline text, @/path/to/file, or - for stdin.

Block until the new participant reaches a terminal state

Wait timeout in seconds (default 300; only with —wait)

Per-POST timeout in seconds for the dispatch call (default 120)

Workspace ID; accepted for consistency (workspace is inferred from <participant_id>)

output the version number

Auth token (or set ISH_TOKEN env var)

Read auth token from a file (preferred over —token / ISH_TOKEN)

Default workspace ID; per-subcommand —workspace overrides

Output as JSON (auto-enabled when piped)

Extract a single field from the JSON response and print only its value (implies —json internally; supports dotted paths e.g. person.name)

Force human-readable output even when stdout is piped (overrides JSON-when-piped auto-detection)

Comma-separated fields to include in JSON output (e.g. alias,name,status)

Include full UUIDs and timestamps in JSON output

Disable colored output (also honored: NO_COLOR env var)

Suppress progress messages on stderr (no-op for read commands that emit none)