Skip to content

Collection Runner

Run a group of saved requests without leaving the terminal interface. You need an initialized collection with requests, and an environment if they use variables. Save pending request and folder changes before starting.

  1. Press F5, or open Ctrl+P and choose Run Collection. F5 resumes live results or the collection’s latest retained run when available.
  2. Select the requests or folders to include. A folder includes its descendants.
  3. Choose the environment. Add Include or Exclude tags to narrow the selection.
  4. Enable Fail-fast if later requests should stop after a failure. Set a delay in milliseconds if the API needs time between requests.
  5. Optionally enter a CSV or JSON Data file path and check the iteration count.
  6. Choose Run to start. Inspect the ordered results as they arrive.

A folder’s context menu opens the Runner scoped to that folder. Selection uses collection order. Within each directory, folders come first, ordered by their meta.seq when set, then by name. Requests follow in filename order. For a sequence such as create then fetch, use filenames like 01-create.yml and 02-fetch.yml. See Collection YAML.

In F4 → Collection → General, choose a Setup request and/or a Teardown request from saved requests. Each is optional; the two must use different request IDs. Settings saves the choices to settings.yml, and the Runner waits for pending collection settings saves before starting.

For each nonempty main selection, Noodle runs setup, the selected requests in collection order, then teardown. With a CSV/JSON data file, this sequence runs once per row. All stages use the normal request lifecycle, including inherited scripts and tests, and share that run or row’s variables and cookie jar. Hook requests stay outside the selectable main list and are not filtered by tags.

A failed setup skips the row’s main requests with reason setup-failed, then attempts teardown. Ordinary failures and Fail-fast also attempt teardown for the started row; fail-fast prevents later data rows from starting. Stop cancels active work and does not start teardown or another request.

Results and details label each stage, and totals include hooks. Settled results keep execution order even when main requests were skipped, along with the hook configuration used for that run. Unassign a hook before renaming or deleting its request, or deleting a containing folder, in the TUI.

See the YAML fields and CLI results for automation.

Request tags and tags inherited from non-root folders form dynamic suites. Include tags must all match; any matching Exclude tag removes a request. Exclusion wins. Tags are case-sensitive. A selection with no matching requests cannot run. See CLI tag examples for the same selection rules in scripts.

Results appear as each request finishes. Rows show pending, running, passed, failed, skipped, or cancelled executions, grouped by iteration when using data. Select a completed row and press Enter, or click it, to inspect the response, script, capture, assertion, and test outcomes while later requests continue. Details include Console when logs are available. Tab and scrolling remain available during execution.

Before HTTP, the TUI Runner checks the syntax of applicable scripts and tests for all selected requests without executing them during validation. With fail-fast enabled, later selected requests are recorded as skipped after all available diagnostics for the failed request finish. A configured teardown is still attempted for the started row unless the run is cancelled.

Live Runner results in an 80 by 24 terminal, with a passed request, a running request, a pending request, and Stop

Press s in the Runner or its completed-result detail overlay, click Stop, or focus the Stop action and press Enter or Space. The command palette also offers Stop Collection Run. The action shows Stopping until shared cancellation and cookie cleanup finish.

Completed results and finished response diagnostics remain available. Cancelled executions stay separate from failures, and selected requests that never start are recorded as skipped because of cancellation. Stopping cannot undo HTTP effects already sent.

Press Escape to leave the Runner while it continues. Read-only browsing, view navigation, and appearance controls remain available. Press F5 to return to the current workspace’s live, partial, or final results.

Request and folder edits, additional sends, collection switches and reloads, and environment, cookie, or transport configuration changes stay locked until the run and cleanup finish. Afterward, use Left from the results list to return to configuration. Editing options does not change the configuration shown with retained results.

Captures and script writes can pass values to later requests in the same run. Each new run starts with a fresh scope. The Runner keeps those values transient even when a request declares persistence.

Each collection keeps its latest settled run in memory until you exit Noodle. Reload the collection, switch to another collection and back, or browse other views, then press F5 to return. The result keeps its original request labels, executed configuration, and selected request/iteration even after you edit the next run’s options or change collection files.

A later settled run replaces the snapshot, including failed or cancelled runs. A configuration or preflight failure leaves the previous result available. Runner snapshots do not create files, edit request declarations, or write Timeline history, and they disappear when the app exits.

Retained text body previews stop at 10,000 UTF-8 bytes and show a truncation notice. Diagnostic limits may omit details; an explicit notice appears while failure summaries and source origins remain available. Binary responses retain metadata only, without a body that can be downloaded from the snapshot.

After the run and cleanup finish, open a completed result’s details and press o, or choose Open source. From Results, this opens the source of the selected script, test, assertion, or capture. Inherited scripts and tests open the owning folder or collection script phase. An aggregate test result with multiple origins offers a searchable picker. From other detail tabs, the action opens the request.

External sources open their declaration; use the existing Ctrl+Alt+X action to open the validated JavaScript file in your configured editor. If a source was renamed or deleted, Noodle warns and keeps the result available. Press F5 to return to the retained result selection. Source navigation is available after cleanup, while completed results remain inspectable during active runs.

Add --dry-run to see the environment and the order of requests for one iteration. For example, preview the smoke-tagged requests under users/ against a CSV file:

noodle collection run ./my-api users/ --tag smoke --data ./data/users.csv --dry-run

With one setup request, two selected requests and one teardown request, each row plans four executions. Three rows give twelve planned executions. Failures can shorten the real run, and scripts may send additional requests. The preview checks the selection and dataset without running scripts, reading secrets or sending HTTP.

Add --json for the ordered sequence of request IDs and stages, the environment name, iterations, and totalExecutions. To save the plan, redirect stdout to a file. The --output and --reporter flags apply to execution reports.

noodle collection run ./my-api --env development

Interrupt CLI collection runs with Ctrl+C to keep partial results and exit 130. See cancellation results.

Use CLI automation and CI for targets, tags, JSON output, exit statuses, and GitHub Actions. Add response assertions to make a run check behavior rather than just send HTTP requests.

noodle collection run ./my-api --data ./data/users.csv
noodle collection run ./my-api users/ --data ./data/users.json --tag smoke --delay 100 --fail-fast --json

--data is optional and independent of the --json output flag. In F5, enter the optional Data file path, check the iteration count, and select Run. Relative CLI paths start at the current directory; relative F5 paths start at the collection root. Both accept absolute paths. The F5 field also completes paths and accepts @/ for your home directory. F5 revalidates at run start. The data-file path used by a settled run stays with its session snapshot; the dataset contents are not retained.

CSV uses a header row as variable names and keeps cells as strings, including quoted commas and multiline fields; a UTF-8 BOM is accepted. JSON requires a non-empty array of objects and preserves JSON value types:

[{ "user_id": 1, "active": true }, { "user_id": 2, "active": false }]

The entire file is validated before requests are sent. Empty files, invalid or unsafe variable names, duplicate CSV headers, and inconsistent CSV rows fail. Names use letters, digits, or underscores. Limits are 5 MiB per file, 1,000 rows, and 10,000 request executions, including configured hooks for each row. Rows must fit the existing 256 KiB/depth-32 bridge limits, including the iteration wrapper.

Each row runs every selected request in collection order. Its variables override environment values for $name substitution and noodle.run.get; successful captures and scripts can replace them during that iteration. noodle.env.get continues to read the original environment snapshot. noodle.iteration is deeply read-only { index, count, data }, with zero-based index and the original row in data. It is null without data and is shared with child requests. JavaScript objects in data retain their JSON types.

Each row starts with fresh variables and an in-memory copy of the same initial cookie jar. Cookies change within that row only and are never persisted. Fail-fast skips remaining main requests and later rows, but still attempts the started row’s teardown unless cancelled. Delay applies between all consecutive executions, including row boundaries. Results, skips, and details carry zero-based iteration; the collection result adds iterations. F5 groups results by iteration and counts progress across all executions. Human labels show iteration numbers starting at one. Datasets themselves are not included in results or history.

Exit codes are 0 for success, 1 for execution/response validation failure, 2 for configuration or invalid data, and 130 for collection cancellation. Without --data, each selected request runs once.