API testing
Explore responses as a tree, assert against a schema, pull secrets from a vault, mask the sensitive bits, run setup and teardown with hooks, and chain it all into visual flows. Everything runs locally first, and the same tests run headless in CI.
Responses render as an expandable tree, so a deeply nested JSON or XML payload is something you navigate instead of scroll. Collapse the noise, drill into the one array that matters, and copy a value or its path in a click.
Click any node to turn it into an assertion. No hand-written path expressions, no guessing whether it is data[0].items[2].price. The tree writes the path for you.
Validate a response against a JSON Schema or an OpenAPI definition and fail the test the moment the shape changes. A renamed field, a string that became a number, a missing required property, all caught automatically rather than three incidents later.
Point a request at a schema once and every run checks it. Contract drift stops being a surprise and becomes a failing assertion with a clear message.
Reference a secret where you need it, for example {{vault:secret/data/oauth#clientId}}, and it is fetched at send time from your provider. The value never gets written into the collection, the environment file, or your git history.
HashiCorp Vault and OpenBao, AWS Secrets Manager, Azure Key Vault, GCP Secret Manager, and 1Password are all supported through the same inline reference syntax.
Define patterns for the values that must never appear in plaintext, and API Spector masks them in the response viewer, the logs, and exported reports. Share a run with a teammate or attach it to a ticket without the token, the card number, or the email going with it.
Masking is applied to live streamed frames too, so what you see on screen matches what leaves the app.
Run logic before and after a request, a folder, or the whole collection with beforeAll, before, after, and afterAll hooks. Seed a record, fetch a token, clean up what you created, all without copying a script into every request.
Pre-request and post-response scripts use a small, documented sp.* API to set variables, read responses, and assert, so setup stays close to the thing it sets up.
Build a visual flow that calls one request, passes a value into the next, branches on a result, and loops over data. A real sequence, drawn on a canvas, instead of a pile of scripts that only the author understands.
Run a flow from the app, headless from the CLI, or in the cloud. The same flow powers a quick manual check and a scheduled run, so there is one definition to maintain.
Fire the request and read the response as a tree. Click the nodes you care about to turn them into assertions.
Attach a schema so drift fails the run, reference secrets from your vault, and mask anything sensitive.
Add setup and teardown hooks, then chain the requests into a flow that models the real sequence.
Run it locally while you build, then headless in CI with the same definition and get a report back.
No. Click a node in the response tree to assert on its value or type. Scripting is available through the sp.* API when you want it, but it is not required for everyday assertions.
JSON Schema and OpenAPI. Attach one to a request and every run checks the response shape, so a renamed or retyped field fails the test instead of slipping through.
No. A vault reference is resolved at send time from your provider and the resolved value is never written to the collection, the environment, or git. Masking keeps it out of logs and reports too.
Yes. Collections and flows run headless from the CLI, so the test you build in the app is the test that runs in your pipeline, with a report you can publish.
The desktop app is free and open source. Add the cloud when you want hosted mocks, monitoring, and a shared broker.