continuedev/continue · error · Error

AI SDK provider requires a model in the format '<provider>/<

Error message

AI SDK provider requires a model in the format '<provider>/<model>' (e.g., 'openai/gpt-4o')

What it means

The AiSdk provider adapter (OpenAI-compatible) derives its provider and model by splitting config.model on '/'; a missing/empty model string makes construction impossible, so the constructor throws immediately.

Source

Thrown at packages/openai-adapters/src/apis/AiSdk.ts:60

      baseURL: options.baseURL ?? "https://openrouter.ai/api/v1/",
    }),
  clawrouter: (options) =>
    createOpenAI({
      ...options,
      baseURL: options.baseURL ?? "http://localhost:1337/v1/",
    }),
};

export class AiSdkApi implements BaseLlmApi {
  private provider?: (modelId: string) => any;
  private config: AiSdkConfig;
  private providerId: string;
  private modelId: string;

  constructor(config: AiSdkConfig) {
    this.config = config;
    if (!config.model) {
      throw new Error(
        "AI SDK provider requires a model in the format '<provider>/<model>' (e.g., 'openai/gpt-4o')",
      );
    }
    const [providerId, ...modelParts] = config.model.split("/");
    this.providerId = providerId;
    this.modelId = modelParts.join("/");
  }

  private initializeProvider() {
    if (this.provider) {
      return;
    }

    const createFn = PROVIDER_MAP[this.providerId];
    if (!createFn) {
      const supportedProviders = Object.keys(PROVIDER_MAP).join(", ");
      throw new Error(
        `Unknown AI SDK provider: "${this.providerId}". ` +

View on GitHub (pinned to 5522c6f44c)

Solutions

1. Set model to '<provider>/<model>' e.g. 'openai/gpt-4o' in the adapter config
2. Check the config source (YAML/config file) for a missing or empty model field
3. Add a startup validation of model strings before constructing providers

Example fix

// before
new AiSdk({ model: '' });

// after
new AiSdk({ model: 'openai/gpt-4o' });

if (!cfg?.model || !/^[^/]+\/.+/.test(cfg.model)) throw new Error('model must be provider/model');

function isProviderModelString(s: unknown): s is `${string}/${string}` { return typeof s === 'string' && /^[^/\s]+\/[^/\s]+$/.test(s); }

try { new AiSdk(config); } catch (e) { if (/requires a model/.test(e.message)) config.model = 'openai/gpt-4o'; else throw e; }

When it happens

Trigger: new AiSdk({ model: '' }) or omitting model in the config passed to the adapter (e.g. via a model config without a 'model' field).

Common situations: YAML/JSON model configs where the model name field is missing or blank, or programmatic config built with a typo'd key.

Related errors

• Either streamCompletion or streamChat must be defined in a c
• No models available in assistant configuration
• Model ${modelName} does not have an apiKeyLocation or envSec
• Unknown AI SDK provider: "${this.providerId}". Supported pro
• Profile ${profileId} not found

AI-assisted analysis of continuedev/continue@5522c6f44c (2026-08-27). Data as JSON: /api/errors/6267c23c75d36e0a. Report an issue: GitHub.