router-for-me/CLIProxyAPI · error

request failed with status %d: %s

Error message

request failed with status %d: %s

What it means

Returned by AntigravityAuth.FetchProjectID when the loadCodeAssist endpoint answers with a non-2xx status. The message includes the status code and the full trimmed response body, so the upstream Google error (usually a JSON block with a 'message' field) is embedded verbatim. Common causes are an invalid/expired OAuth access token (401), missing cloudaicompanion API entitlements (403), or a malformed request (400).

Source

Thrown at internal/auth/antigravity/auth.go:264

	req.Header.Set("User-Agent", userAgent)

	resp, errDo := o.httpClient.Do(req)
	if errDo != nil {
		return "", fmt.Errorf("execute request: %w", errDo)
	}
	defer func() {
		if errClose := resp.Body.Close(); errClose != nil {
			log.Errorf("antigravity loadCodeAssist: close body error: %v", errClose)
		}
	}()

	bodyBytes, errRead := io.ReadAll(resp.Body)
	if errRead != nil {
		return "", fmt.Errorf("read response: %w", errRead)
	}

	if resp.StatusCode < http.StatusOK || resp.StatusCode >= http.StatusMultipleChoices {
		return "", fmt.Errorf("request failed with status %d: %s", resp.StatusCode, strings.TrimSpace(string(bodyBytes)))
	}

	var loadResp map[string]any
	if errDecode := json.Unmarshal(bodyBytes, &loadResp); errDecode != nil {
		return "", fmt.Errorf("decode response: %w", errDecode)
	}

	projectID := extractCloudaicompanionProject(loadResp)

	if projectID == "" {
		projectID, err = o.OnboardUser(ctx, accessToken, defaultAntigravityTierID(loadResp))
		if err != nil {
			return "", err
		}
		if projectID == "" {
			return "", fmt.Errorf("project id not found in loadCodeAssist or onboardUser response")
		}
		return projectID, nil

View on GitHub (pinned to 78f0c4079e)

Solutions

1. Read the embedded body: a 401 with 'Invalid Credentials' means the access token is stale — re-run the antigravity login flow to mint fresh credentials
2. A 403 with 'permission denied' on cloudaicompanion.googleapis.com means the account lacks entitlement — enable Gemini code assist or use an account that has it
3. A 5xx is transient: wait and retry the auth flow
4. Verify the system clock; skewed clocks can invalidate freshly issued tokens

if strings.TrimSpace(accessToken) == "" {
    return fmt.Errorf("refusing loadCodeAssist with empty token; refresh first")
}

_, err := auth.FetchProjectID(ctx, token)
if err != nil && strings.Contains(err.Error(), "request failed with status") {
    if strings.Contains(err.Error(), "401") {
        // token stale: force re-login
    } else if strings.Contains(err.Error(), "403") {
        // account not entitled: surface to user
    } else {
        // 5xx: retry later
    }
}

When it happens

Trigger: Calling FetchProjectID with an access token that expired between refresh and this call (401); a Google account without Gemini/Antigravity code-assist access enabled (403); upstream 5xx during a Google incident.

Common situations: Stale token file under auths/ where the refresh token flow succeeded but the new token was rejected; GoogleWorkspace accounts with the Gemini for Workspace app policy disabled; regional blocks on cloudcode-pa.googleapis.com.

Related errors

• antigravity token exchange: request failed: status %d
• antigravity token exchange: request failed: status %d: %s
• antigravity userinfo: request failed: status %d
• antigravity userinfo: request failed: status %d: %s
• http %d: %s

AI-assisted analysis of router-for-me/CLIProxyAPI@78f0c4079e (2026-08-15). Data as JSON: /api/errors/e0d0c259c815ad64. Report an issue: GitHub.