Service twin · API + UI

Check

Test payroll agents and integrations that onboard workers, run payroll, and coordinate payments and filings.

API + UIdocs-transcribedREST · MCP · CLI

Start a twin run Read the quickstart Open the live explorer

Terminal

$ uv tool install blobfish-cli
$ blobfish twin login
$ blobfish twin-runs create --twins check --ttl 60 --wait

What’s inside

A deterministic, seeded, resettable replica: every operation below executes real state transitions against the twin’s own SQLite state — reads are hash-invariant, writes produce inspectable diffs, and reset restores the seed baseline.

Operations31

Writes17

Entities7

Lifecycles3

Seeded entities

Entity	Table	Fields	About
company	`companies`	7	Employer of record that runs payroll through Check.
workplace	`workplaces`	8	Physical work location used for tax jurisdiction assignment.
employee	`employees`	9	W-2 employee with an onboarding lifecycle that must complete before payroll.
contractor	`contractors`	8	1099 contractor paid outside the W-2 payroll flow.
payroll	`payrolls`	9	Pay run for a company: drafted, approved into pending, processed on payday, then paid.
payroll_item	`payroll_items`	8	One employee's earnings line inside a payroll run.
tax_filing	`tax_filings`	8	Tax filing/payment obligation Check prepares and files with a jurisdiction.

Guarded lifecycles

These state machines are enforced, not decorative: an action whose fromstate doesn’t match is rejected the way the production service rejects it.

employee.onboard_status

blocking → needs_attention → completed

complete_onboarding: blocking|needs_attention → completed

payroll.status

draft → pending → processing → paid → failed

approve: draft → pending
reopen: pending → draft

tax_filing.status

draft → submitted → accepted → rejected

file: draft → submitted

Three surfaces, one state

Mount Check in a twin environment and the same seeded state serves vendor-shaped REST operations, an MCP endpoint for tool calling, and a generated bf-twin CLI — with copy-on-write sessions per rollout and a one-call reset to the seed baseline. Prefer raw HTTP? The same create works with curl:

API

$ curl -sS -X POST https://blobfish.ai/api/v1/twin/environments \
    -H "X-API-Key: $BLOBFISH_API_KEY" -H "Content-Type: application/json" \
    -d '{"services":["check"],"ttl_minutes":60}'

Known limitations

transcribed from https://docs.checkhq.com/reference; parameter and response schemas simplified
nested Check resources flattened to top-level collections addressed by *_id filters; employee onboarding completion modeled as an explicit action
synthetic seeded values