n8n-io/n8n · error · NodeOperationError

Invalid authentication method

Error message

Invalid authentication method

What it means

The default branch of a switch on authenticationMethod in the Azure OpenAI Chat node. The switch only handles AuthenticationType.ApiKey and AuthenticationType.EntraOAuth2; any other value (or an unknown future auth type) hits the default and throws NodeOperationError before the client is built.

Source

Thrown at packages/@n8n/nodes-langchain/nodes/llms/LmChatAzureOpenAi/LmChatAzureOpenAi.node.ts:98

				itemIndex,
			) as AuthenticationType;
			const modelName = this.getNodeParameter('model', itemIndex) as string;
			const options = this.getNodeParameter('options', itemIndex, {}) as AzureOpenAIOptions;

			// Set up Authentication based on selection and get configuration
			let modelConfig: AzureOpenAIApiKeyModelConfig | AzureOpenAIOAuth2ModelConfig;
			switch (authenticationMethod) {
				case AuthenticationType.ApiKey:
					modelConfig = await setupApiKeyAuthentication.call(this, 'azureOpenAiApi');
					break;
				case AuthenticationType.EntraOAuth2:
					modelConfig = await setupOAuth2Authentication.call(
						this,
						'azureEntraCognitiveServicesOAuth2Api',
					);
					break;
				default:
					throw new NodeOperationError(this.getNode(), 'Invalid authentication method');
			}

			this.logger.info(`Instantiating AzureChatOpenAI model with deployment: ${modelName}`);

			const timeout = options.timeout;
			const model = new AzureChatOpenAI({
				// Force completions API — Azure's SDK doesn't rewrite the /responses path,
				// so the Responses API hits an invalid endpoint and causes a connection error.
				// See: https://github.com/langchain-ai/langchainjs/issues/9038
				useResponsesApi: false,
				// Model name is required so logs are correct
				// Also ensures internal logic (like mapping "maxTokens" to "maxCompletionTokens") is correct
				model: modelName,
				azureOpenAIApiDeploymentName: modelName,
				...modelConfig,
				...options,
				timeout,
				maxRetries: options.maxRetries ?? 2,

View on GitHub (pinned to 5ac6606e81)

Solutions

1. Open the node credential selector and re-pick 'API Key' or 'Entra ID (OAuth2)'.
2. Recreate the credential if the dropdown does not show a valid selection.
3. Upgrade n8n / nodes-langchain so the auth type enum matches the code's switch arms.

Example fix

// before
authenticationMethod === 'azureActiveDirectory' // no longer handled
// after
authenticationMethod === 'entraOAuth2'

const ALLOWED = [AuthenticationType.ApiKey, AuthenticationType.EntraOAuth2];
if (!ALLOWED.includes(authenticationMethod)) {
  throw new Error(`Pick 'API Key' or 'Entra ID (OAuth2)' in the credential selector.`);
}

const isSupportedAuth = (m: unknown): m is AuthenticationType =>
  m === AuthenticationType.ApiKey || m === AuthenticationType.EntraOAuth2;

// Validate the enum at node-save time; surface allowed values.

When it happens

Trigger: authenticationMethod parameter holds a value outside {apiKey, entraOAuth2} — e.g. from a hand-edited workflow JSON, a stale value from a deprecated auth type, or a forward-incompatible enum.

Common situations: Importing an older workflow that used a removed auth method; manually editing the workflow JSON; future n8n version adding a new auth type that the current code doesn't handle; corrupted credential parameter.

Understand the failure class

• Authentication and authorization failures — expired tokens, bad credentials, and missing scopes.

Related errors

• The value in the "Extra Body" field is not valid JSON
• The value in the "Extra Body" field must be a JSON object
• Failed to initialize Azure OpenAI client: ${error.message}
• API Key is missing in the selected Azure OpenAI API credenti
• VectorStore "${this.name}" requires a description — set it v

AI-assisted analysis of n8n-io/n8n@5ac6606e81 (2026-08-12). Data as JSON: /api/errors/07d33c8c7b7ef476. Report an issue: GitHub.