Skip to main content
A card above the chat box asking for the bearer token for Staging: the token field, Save and Not Now

How It Works

For an API project, an environment is your API’s base address plus how to authenticate against it. The agent sets that up with you in the chat:
1

Tell the agent how your API authenticates

Say it in plain words (“Bearer token in the Authorization header”, “POST /auth/login with email and password, the token is at data.access_token”), or attach your API docs and let it read them.
2

Confirm the environment

The agent writes the environment with everything except the secrets. A card asks: Save N changes to environments? Click Save.
3

Type each secret on its card

For every secret the environment needs, a card named for it appears with a password field. Type the value and click Save.
4

The agent checks it

For a sign-in flow, saving the secret tries the sign-in once, and the agent hears whether it worked. It then calls your API with the credential in place.

Secrets Never Go Through the Chat

A secret typed on a card is saved straight to the environment, encrypted. It never appears in the chat, and the agent never sees it: it learns only whether the secret was saved and, for a sign-in, whether signing in works now.
Don’t paste a token, key or password into a message. If you do, a card stops it before it is sent: “This looks like a key or a password.” Click Edit Message, remove it, and give it on the secret card instead.

Supported Auth Types

For Login, Refresh token and AWS Cognito, the agent also sets where the fresh token goes on each request: as a bearer token, in a named header, or in a named cookie. These three are Auto-Auth flows, so TestSprite fetches a fresh token before every run, and they follow Auto-Auth’s plan availability.

Headers

An environment can also carry headers sent with every request, such as x-org-id. The agent sets ordinary header values itself. A header whose name looks like a secret (a token, key, secret, auth, session, password, signature or cookie) is treated as one: its value is collected on a card and shows only as set or unset afterwards.

When a Credential Is Wrong or Expires

  • If a call comes back 401 or 403, the agent checks the auth setup and, when the secret is the problem, asks for it again with a new card.
  • If a sign-in flow fails, the environment records the error, and the agent reads it to work out what to fix.
  • You can also change a credential yourself: open the environment from the Setup tab, or go to Settings → Environment. See Environments.
Give TestSprite a long-lived credential for a dedicated test account with access to the whole API. The agent’s calls and every test run use it.

More Than One Environment

Each environment has its own base address and its own auth. The agent works on the project’s default environment unless you name another, and a run’s card shows the environment it will use and lets you pick a different one.
The plan card's arrow open beside Apply and Run: 1 the arrow, 2 Run on Staging or Run on Production

Troubleshooting

Secrets are asked for after the environment is saved. Confirm the Save N changes to environments? card first; only one card waits at a time.
Check the auth type matches how your API reads the credential (a bearer token sent as an API key won’t work), then ask the agent to ask for the secret again.
The token’s path in the response is probably wrong. Tell the agent where it is (“the token is at data.session.jwt”) and it updates the environment.
These aren’t auth types the agent can set up. Tell the agent what your API requires; it will say what it can and can’t cover.

Where to Go Next

API Calls

How the agent uses the credential to learn and test your API

Auto-Auth

Keep tokens fresh on every run

Environments

Addresses, credentials and headers per environment

Setup (UI)

The web-app equivalent: test accounts and saved sign-ins