Execute API

Errors

How the Execute API says what went wrong, and what it would accept instead.

Shape

Every error has the same shape: a stable code, a message for people, and, when the body was the problem, every issue at once.

400 Bad Request

{
  "error": {
    "code": "invalid_intent",
    "message": "/chainId: expected a JSON integer such as 8453; received the string \"8453\"",
    "issues": [
      {
        "path": "/chainId",
        "message": "expected a JSON integer such as 8453; received the string \"8453\""
      }
    ]
  }
}

code is stable: branch on it.
message is for people, and may change.
issues holds one entry per problem, each with a path (a JSON Pointer into your body) and a message that says what would be accepted. It’s empty when the body wasn’t the problem.

Codes

Status	Code	When
400	`invalid_intent`	The intent isn’t valid. issues lists every problem.
400	`invalid_request`	Anything else malformed: a path that isn’t an address, both credentials at once.
401	`unauthorized`	No credential, or an unknown one.
403	`api_key_in_browser`	An API key sent from a browser.
403	`origin_required`	A client ID sent from outside a browser.
403	`origin_not_allowed`	A client ID sent from an origin the account doesn’t allow.
404	`not_found`	Execute has no intent for the address.
409	`address_conflict`	The same calldata was registered with other function signatures.
429	`rate_limited`	Over the credential’s rate limit.
503	`unavailable`	Execute can’t serve the request right now. Retry with backoff.
500	`internal`	Something failed on Execute’s side. Retry with backoff.

Every issue Create an address reports is listed with it.

Request ids

Every response carries an X-Request-Id header. Include it when you ask about a request, and it can be found in Execute’s logs.