Multimeter is a Git-native REST client and API testing tool for VS Code. Requests, tests, mocks, and docs are YAML .mmt files in Git — the same files in the visual editor and in CI via testlight. Apache 2.0. No account.

How do I use Multimeter with Cursor, Copilot, or Claude?

Install the VS Code/Cursor extension or add MCP with npx -y mmt-mcp. Agents should call scaffold_test, validate, format, and run instead of guessing .mmt syntax.

Yes. Apache License 2.0. No account required.

Inline expect — Multimeter docs

Use expect / require on a call or http step to validate outputs inline, without a separate check/assert step. Each key is an output field name; each value…Markdown (no JavaScript)# Inline expect, require, and debug Use `expect` / `require` on a [`call`](./steps/call.md) or [`http`](./steps/http.md) step to validate outputs inline, without a separate `check`/`assert` step. Each key is an output field name; each value is the expected result. Soft (`expect`) and hard (`require`) entries on the same call share **one report item**. | Block | On failure | |-------|------------| | `expect` | Log and report; **continue** (like [check](./steps/check.md)) | | `require` | Log and report; **stop** (like [assert](./steps/assert.md)) | Full `call` field reference: [call](./steps/call.md). **Formats:** ```yaml # Soft - call: login expect: status_code: 200 # Hard - call: login require: token: != null # Soft + hard (one report box) - call: login expect: status_code: 200 require: token: != null ``` Expected sides follow the same YAML typing as `if` / `check`: unquoted `200` / `true` / `null` are number / bool / null; quote to force a string (`== "200"`). Write **all** operator expects **unquoted** — `!= null`, `>=# 24`, `=* …`, `=1d~ <<c:datetime>>`, etc. Do not hand-quote them (`'>=# 24'` is wrong); Multimeter handles YAML-unsafe ops. Prefer **accurate** expects over bare `!= null`: live timestamps with `=Ns~ <<c:datetime>>` / `<<c:date>>` (not `c:time`), fixed windows with `=^` / exact `e:` dates, numeric ranges, string contains/prefix, length (`=#` / `>=#`), fuzzy text (`>N%`), or exact equality when the value is known. See [Operators — Prefer accurate expects](./operators.md#prefer-accurate-expects-not-only--null). **Prefer `expect` / `require` over a separate `check` / `assert` whenever the comparison only uses that call’s outputs** (plus literals / `e:` env). Keep standalone [check](./steps/check.md) / [assert](./steps/assert.md) for cross-step or two-output comparisons. Set `report: all` on the call when you want every expect row shown in the Report list (default external reporting is often fails-only). ```yaml # Multiple expects on the same field (not follow-up - check:) - call: login expect: status_code: - == 200 - != 500 report: all # Nested field access - call: getUser expect: body.user.name: == John body.user.active: true report: all # With title and report - call: login title: Login validation expect: status_code: 200 token: != null report: all ``` All comparison operators from [Operators](./operators.md) are available in `expect` / `require` values. Missing fields use [omit](./omit.md). Report levels (`all` / `fails` / `none`): [Reports — Report levels](./reports.md#report-levels-all-fails-none). #### Inline debug `debug` uses the same syntax and [operators](./operators.md) as `expect`, but results show a **debug icon** and are **not included in exported reports**. ```yaml - call: login expect: status_code: 200 debug: body.token: != null body.expires_in: > 0 ``` See also: [Operators](./operators.md) · [omit](./omit.md) · [call](./steps/call.md) · [check](./steps/check.md) · [run](./steps/run.md) · [Reports](./reports.md#report-levels-all-fails-none) For AI agents · llms.txt · Home