CLI

Frank ships two CLI surfaces. They solve different problems and should both stay documented.

Tool	Location	Audience	Purpose
`frankctl`	`frank-cli/`	Operators, developers, CI, agent runtimes	Operate the Frank API: auth, config, sources, streams, transforms, pipelines, datasets, schedules, runs, and AI workflows.
`frank`	`utils/frank_cli/`	Transform pattern authors	Scaffold, validate, test, and build custom transform pattern containers.

Use frankctl to control the platform. Use frank when you are building reusable transform pattern packages.

`frankctl`

frankctl is the TypeScript, terminal-first API client.

bash

cd frank-cli
npm install
npm run build
npm link
frankctl --help

For local Compose, the API is usually exposed on 8002:

bash

frankctl --api-url http://localhost:8002 status
frankctl config set profiles.default.apiUrl http://localhost:8002

Global options

Available on every command:

Flag	Purpose
`-p, --profile <name>`	Select config profile.
`--json`	Emit machine-readable JSON when the command supports table output.
`-v, --verbose`	Verbose logging.
`-q, --quiet`	Suppress non-error output.
`--no-color`	Disable ANSI colors.
`--api-url <url>`	Override the API base URL for this invocation.

Environment variables

Variable	Purpose
`FRANKCTL_API_URL`	API base URL override.
`FRANKCTL_KEYCLOAK_URL`	Keycloak base URL.
`FRANKCTL_KEYCLOAK_REALM`	Keycloak realm.
`FRANKCTL_CLIENT_ID`	OAuth client ID.
`FRANKCTL_CLIENT_SECRET`	Service account client secret.
`FRANKCTL_PROFILE`	Active config profile.
`FRANKCTL_TOKEN`	Raw JWT override; bypasses token store.
`FRANK_DEV_MODE=true`	Enables `X-Tenant-ID` forwarding for local development.
`FRANKCTL_TENANT_ID`	Tenant ID forwarded when `FRANK_DEV_MODE=true`.
`PATTERN_WEBHOOK_SECRET`	HMAC key for `frankctl patterns register`.

Exit codes

Code	Meaning
`0`	Success.
`1`	Generic error.
`2`	Usage error.
`3`	Authentication or authorization error.
`4`	API or service error.
`5`	Validation failed.
`6`	CLI config missing or invalid.

Config and status

bash

frankctl status
frankctl status --json

frankctl config set profiles.default.apiUrl http://localhost:8002
frankctl config get apiUrl
frankctl config list

config get accepts effective profile fields:

text

apiUrl
keycloakUrl
realm
clientId
clientSecret

Config precedence is command flag, environment, config file, then built-in defaults.

Authentication

bash

# Browser PKCE
frankctl auth login

# Headless password grant
frankctl auth login --headless --username you@example.com --password '...'

# Service account
frankctl auth login --service-account \
  --client-id frank-api \
  --client-secret "$FRANKCTL_CLIENT_SECRET"

frankctl auth status
frankctl auth token
frankctl auth logout

Tokens are stored per profile. frankctl auth token prints a valid access token to stdout for use in scripts.

Sources

Manage source connections and source syncs.

bash

frankctl sources list [--type <pattern-id>] [--status <s>] [--search <q>] [--page N] [--page-size N]
frankctl sources get <id> [--no-streams]
frankctl sources create -f source.yaml [--name <name>] [--type <pattern-id>]
frankctl sources update <id> -f patch.yaml
frankctl sources delete <id> --yes

Discovery:

bash

frankctl sources discover <id> [--no-wait] [--timeout 300] [--poll-interval 2000]

Sync:

bash

frankctl sources sync <id> \
  [--streams customers,orders] \
  [--force-full-refresh] \
  [--no-wait] \
  [--timeout 1800] \
  [--poll-interval 3000]

Logs and history:

bash

frankctl sources history <id> [--limit 20] [--offset 0] [--status <s>]
frankctl sources logs <id> <run-id> [-f] [--level <level>] [--poll-interval 2000]

sources create and sources update read YAML, JSON, or - for stdin.

Source streams

bash

frankctl sources streams list <source-id> [--enabled-only]
frankctl sources streams set <source-id> -f streams.yaml
frankctl sources streams refresh-schema <source-id> --from-discovery [--timeout 300] [--poll-interval 2000]
frankctl sources streams refresh-schema <source-id> -f discovery.json

streams set accepts either a bare stream array or the schema accepted by the CLI's bulk stream validator. refresh-schema either consumes a saved discovery payload or runs discovery first.

Transforms

bash

frankctl transforms list [--status <s>] [--search <q>] [--pipeline-id <id>] [--page N] [--page-size N]
frankctl transforms get <id>
frankctl transforms runs <id> [--status <s>] [--limit 50] [--page N]
frankctl transforms logs <id> <run-id> [-f] [--level <level>] [--poll-interval 2000]
frankctl transforms trigger <id>

Use transforms trigger for Run Now. It posts to:

http

POST /api/v1/transforms/{id}/schedule/trigger

That endpoint is the transform "Run Now" path. It dispatches through Dagster, even for manual schedules, and returns the Dagster run information.

Pipelines

bash

frankctl pipelines list [--status <status>] [--search <q>] [--page N] [--page-size N]
frankctl pipelines get <id> [--include-version]
frankctl pipelines validate <id> [--sample-limit 1000] [--timeout 600] [--poll-interval 2500] [--no-wait]

pipelines validate starts a sandbox run and polls until terminal status unless --no-wait is supplied. It always emits final JSON to stdout. Step progress is written to stderr.

Runs

bash

frankctl runs get <workflow-id>
frankctl runs wait <workflow-id> [--timeout 600] [--poll-interval 2000]
frankctl runs cancel <id> --yes

For UUID IDs, runs cancel tries transform-run cancellation first and then source sync-run cancellation. For non-UUID IDs, it calls Temporal workflow cancellation.

Patterns

Pattern commands cover the platform's pattern catalog and external pattern lifecycle. Listing and detail commands browse source patterns; parameter validation checks transform pattern params; registration is used by CI publishing flows.

Commands:

bash

frankctl patterns list [--category <c>] [--search <q>]
frankctl patterns get <id>
frankctl patterns validate-params <pattern-id> -f params.yaml
frankctl patterns register -f webhook-payload.json

patterns register is CI-oriented and requires PATTERN_WEBHOOK_SECRET. The CLI signs the payload with X-Hub-Signature-256.

Schedules

The frankctl schedules tree is source-scoped.

bash

frankctl schedules list
frankctl schedules get <source-id>
frankctl schedules set <source-id> -f schedule.yaml
frankctl schedules pause <source-id>
frankctl schedules resume <source-id>
frankctl schedules trigger <source-id>
frankctl schedules delete <source-id> --yes

Example:

yaml

schedule_type: cron
schedule_value: "0 */6 * * *"

For transform execution, use frankctl transforms trigger <id>.

Datasets

bash

frankctl datasets list [--layer bronze|silver|gold] [--namespace <ns>] [--page N] [--page-size N]
frankctl datasets preview <layer.namespace.table> [--limit 20]
frankctl datasets snapshots <layer.namespace.table>

Dataset IDs use the dotted three-part form accepted by the FastAPI {dataset_id:path} route.

AI

All AI commands call Frank's Martha-backed AI routes and emit JSON.

bash

frankctl ai compose-pipeline -f spec.yaml [--no-wait] [--timeout 180] [--poll-interval 2500]
frankctl ai generate-transform -f spec.yaml
frankctl ai review-sql -f spec.yaml
frankctl ai fix-ci-failure -f spec.yaml
frankctl ai suggest target-schema -f spec.yaml
frankctl ai suggest field-mappings -f spec.yaml
frankctl ai suggest pattern-params -f spec.yaml
frankctl ai publish-transform -f spec.yaml

compose-pipeline starts an execution and polls /api/v1/ai/compose-pipeline/{execution_id}/status unless --no-wait is set. If the backend returns no execution_id, the CLI treats it as validation failure and does not poll.

Python `frank` pattern CLI

The Python frank CLI is part of utils/ and installs from the frank-shared package.

bash

cd utils
pip install -e ".[cli]"
frank --help

This tool is for authoring custom transform patterns. It does not manage deployed sources, transforms, pipelines, or runs.

Scaffold a pattern

bash

frank init <name> [--output patterns] [--template basic|sql|python]

Examples:

bash

frank init my-filter
frank init my-aggregator --template sql
frank init my-transform --output ./custom-patterns

Creates:

text

pattern.yaml
main.py
requirements.txt
Dockerfile
tests/test_transform.py
tests/sample_config.json
README.md

Validate a pattern

bash

frank validate [path] [--strict]

Validation checks:

pattern.yaml exists.
Required schema fields are present.
Field types match the pattern schema.
Pattern ID format is valid.
params_schema is structurally sensible.
Container image hints are reasonable.
main.py exists.

--strict fails on warnings as well as errors.

Test a pattern locally

bash

frank test [path] [--config tests/sample_config.json] [--env-file .env.test]

The test command loads a TRANSFORM_CONFIG-shaped JSON file, applies optional env vars, runs main.py, parses the emitted FrankResult, and can run Trino fixture setup/assertion/teardown blocks from the sample config.

Example config:

json

{
  "artifact_id": "test-local-001",
  "run_id": "test-run-001",
  "tenant_id": "dev-tenant",
  "source_table": "iceberg.bronze.orders",
  "target_table": "iceberg.silver.orders_clean",
  "params": {
    "filter_expression": "status != 'deleted'"
  }
}

Build a pattern image

bash

frank build [path] --tag <image-tag> [--push] [--registry ghcr.io/colivetree]

Examples:

bash

frank build patterns/my-filter --tag my-filter:v1
frank build . --tag ghcr.io/acme/my-filter:v1 --push

If the pattern directory has no Dockerfile, the CLI generates one from pattern.yaml.

Which tool should I use?

Task	Tool
Log in to Frank	`frankctl auth login`
Create or sync a source	`frankctl sources ...`
Configure source streams	`frankctl sources streams ...`
Trigger a transform run	`frankctl transforms trigger <id>`
Validate a pipeline sandbox	`frankctl pipelines validate <id>`
Preview an Iceberg dataset	`frankctl datasets preview ...`
Ask AI to compose or review	`frankctl ai ...`
Scaffold a custom transform pattern	`frank init ...`
Validate a local pattern package	`frank validate ...`
Run a pattern locally	`frank test ...`
Build and push a pattern image	`frank build ...`

CLI ​

frankctl ​

Global options ​

Environment variables ​

Exit codes ​

Config and status ​

Authentication ​

Sources ​

Source streams ​

Transforms ​

Pipelines ​

Runs ​

Patterns ​

Schedules ​

Datasets ​

AI ​

Python frank pattern CLI ​

Scaffold a pattern ​

Validate a pattern ​

Test a pattern locally ​

Build a pattern image ​

Which tool should I use? ​

CLI

`frankctl`

Global options

Environment variables

Exit codes

Config and status

Authentication

Sources

Source streams

Transforms

Pipelines

Runs

Patterns

Schedules

Datasets

AI

Python `frank` pattern CLI

Scaffold a pattern

Validate a pattern

Test a pattern locally

Build a pattern image

Which tool should I use?