n8n-io/n8n · error · Error

Unsupported provider: "${provider}". Supported providers: ${

Error message

Unsupported provider: "${provider}". Supported providers: ${SUPPORTED_PROVIDERS}

What it means

After splitting the model ID at the first `/`, `createModel` looks up the provider segment in `LANGUAGE_PROVIDERS`. If the provider key is not in the registry it throws, listing all supported providers. The registry currently includes: `openai`, `custom`, `anthropic`, `google`, `google-vertex`, `mistral`, `cohere`, `microsoft-azure`, `aws-bedrock` (per the registry definition).

Source

Thrown at packages/@n8n/agents/src/runtime/model/model-factory.ts:239

	if (isLanguageModel(config)) {
		return config;
	}

	const rawId = typeof config === 'string' ? config : config.id;
	if (!rawId || rawId.trim() === '') {
		throw new Error('Model ID is required');
	}

	const slashIndex = rawId.indexOf('/');
	if (slashIndex <= 0) {
		throw new Error(`Invalid model ID "${rawId}": expected "provider/model-name" format`);
	}
	const provider = rawId.slice(0, slashIndex) as ProviderId;
	const modelName = rawId.slice(slashIndex + 1);

	const entry = LANGUAGE_PROVIDERS[provider];
	if (!entry) {
		throw new Error(
			`Unsupported provider: "${provider}". Supported providers: ${SUPPORTED_PROVIDERS}`,
		);
	}

	// Collect credential fields: strip `id`, pass the rest to Zod validation.
	let credFields: Record<string, unknown> = {};
	if (typeof config !== 'string') {
		const { id: _id, ...rest } = config as { id: string; [k: string]: unknown };
		credFields = rest;
	}
	// Host configs (e.g. Instance AI's `{ id, url }` for OpenAI-compatible
	// endpoints) spell the base URL as `url`; the provider schemas only know
	// `baseURL`, and Zod strips unknown keys, so normalize before validation.
	// An EMPTY url means "no custom endpoint" (Instance AI emits `url: ''` for
	// the api-key-only config) and must keep the provider default.
	if (typeof credFields.url === 'string' && credFields.baseURL === undefined) {
		const { url, ...restCreds } = credFields;
		credFields = url ? { ...restCreds, baseURL: url } : restCreds;

View on GitHub (pinned to 5ac6606e81)

Solutions

1. Read the error message — it lists all supported provider keys.
2. Correct the provider segment to match a registered key exactly (e.g. `anthropic` not `anthrpic`).
3. For OpenAI-compatible custom endpoints, use the `custom/` provider prefix and supply `baseURL` and `apiKey` in the config.
4. If you need a genuinely new provider, add an entry to `LANGUAGE_PROVIDERS` in model-factory.ts and install the corresponding `@ai-sdk/*` package.

Example fix

// before:
createModel('anthrpic/claude-sonnet-4-5'); // typo

// after:
createModel('anthropic/claude-sonnet-4-5');

import { SUPPORTED_PROVIDERS } from './model-factory'; // export if needed

const KNOWN_PROVIDERS = new Set(['openai','custom','anthropic','google','google-vertex','mistral','cohere','microsoft-azure','aws-bedrock']);

function validateProvider(modelId: string): void {
  const provider = modelId.slice(0, modelId.indexOf('/'));
  if (!KNOWN_PROVIDERS.has(provider)) {
    throw new Error(`Unknown provider "${provider}". Known: ${[...KNOWN_PROVIDERS].join(', ')}`);
  }
}

validateProvider(config.modelId);

function isKnownProvider(modelId: string): boolean {
  const provider = modelId.slice(0, modelId.indexOf('/'));
  return ['openai','custom','anthropic','google','google-vertex','mistral','cohere','microsoft-azure','aws-bedrock'].includes(provider);
}

When it happens

Trigger: Calling `createModel('huggingface/llama-3')` where `huggingface` is not a registered provider. Using a typo like `createModel('anthrpic/claude-sonnet-4-5')`. Using a provider alias that does not exist in the registry.

Common situations: Typo in the provider name. The desired provider's `@ai-sdk/*` package is not installed (though this error is about the registry key, not the package). A custom/self-hosted model endpoint was given a provider name not in the registry (should use `custom/` instead).

Related errors

• Unsupported embedding provider: "${provider}". Supported: ${
• Model ID is required
• Invalid model ID "${rawId}": expected "provider/model-name"
• Invalid credentials for provider "${provider}": ${issues}
• Cannot decrease maxIterations when resuming a run. Expected

AI-assisted analysis of n8n-io/n8n@5ac6606e81 (2026-08-12). Data as JSON: /api/errors/60b05676274d6e0f. Report an issue: GitHub.