Manage Ascend runtimes, flows, and flow runs from the command line.
uv tool install ascend-toolsUpgrade to the latest version:
uv tool install --upgrade ascend-toolsSee Installation for other methods (Cargo, pre-built binaries).
Set three environment variables (see Quickstart for the full service account creation walkthrough):
export ASCEND_SERVICE_ACCOUNT_ID="<YOUR_SERVICE_ACCOUNT_ID>"
export ASCEND_SERVICE_ACCOUNT_KEY="<YOUR_SERVICE_ACCOUNT_KEY>"
export ASCEND_INSTANCE_API_URL="<YOUR_INSTANCE_API_URL>"You can also pass credentials as CLI flags: --service-account-id, --service-account-key, --instance-api-url. Flags override environment variables.
ascend-tools runtime listFilter by ID, kind, project, or environment:
ascend-tools runtime list --id my-runtime
ascend-tools runtime list --kind deployment
ascend-tools runtime list --project-uuid <UUID>
ascend-tools runtime list --environment-uuid <UUID>ascend-tools runtime get <RUNTIME_UUID>ascend-tools runtime pause <RUNTIME_UUID>ascend-tools runtime resume <RUNTIME_UUID>ascend-tools flow list --runtime <RUNTIME_UUID>ascend-tools flow run <FLOW_NAME> --runtime <RUNTIME_UUID>Resume a paused runtime before running:
ascend-tools flow run <FLOW_NAME> --runtime <RUNTIME_UUID> --resumePass a flow run spec for advanced options:
ascend-tools flow run <FLOW_NAME> --runtime <RUNTIME_UUID> \
--spec '{"full_refresh": true}'| Field | Type | Description |
|---|---|---|
full_refresh |
bool | Drop all internal data and recompute from scratch. Destructive. |
components |
list | Run only these components (by name). Omit to run all. |
component_categories |
list | Run only components in these categories. |
parameters |
object | Custom parameters passed to the flow. |
run_tests |
bool | Run tests after processing data. Defaults to true. |
store_test_results |
bool | Store test results. |
halt_flow_on_error |
bool | Stop the flow on error. |
disable_optimizers |
bool | Disable optimizers. |
update_materialization_type |
bool | Update component materialization types. May drop data. |
deep_data_pruning |
bool | Full table scan for Smart Table data maintenance. |
backfill_missing_statistics |
bool | Backfill statistics for data blocks without them. |
disable_incremental_metadata_collection |
bool | Disable incremental read/transform metadata collection. |
runner_overrides |
object | Runner config overrides (e.g., {"size": "Medium"}). |
ascend-tools flow list-runs --runtime <RUNTIME_UUID>Filter by status, flow name, time range, or paginate:
ascend-tools flow list-runs --runtime <RUNTIME_UUID> --status running
ascend-tools flow list-runs --runtime <RUNTIME_UUID> --flow-name sales
ascend-tools flow list-runs --runtime <RUNTIME_UUID> --since 2025-01-01T00:00:00Z
ascend-tools flow list-runs --runtime <RUNTIME_UUID> --limit 10 --offset 20ascend-tools flow get-run <RUN_NAME> --runtime <RUNTIME_UUID>Default output is a human-readable table. Use -o json for machine-readable output:
ascend-tools -o json runtime list
ascend-tools -o json flow list-runs --runtime <RUNTIME_UUID>Empty results print "No results." to stderr.
Install reference skills for AI coding assistants (Claude Code, Codex, etc.):
ascend-tools skill install --target .claude/skills --allAvailable flags: --cli (default), --python, --mcp, --all.