continuedev/continue · error · Error

Unsupported model

Error message

Unsupported model: ${body.model}

What it means

The embed method dispatches by model family: only known Bedrock embedding models (e.g. Amazon Titan / Cohere embed models the adapter recognizes) are handled; any other model string falls into the else branch and throws `Unsupported model: <model>`. It is a whitelist-based capability gate, not an AWS-side rejection.

Solutions

1. Check the supported embedding model list for your adapter version and use one of those exact ids (e.g. amazon.titan-embed-text-v2:0).
2. Upgrade the openai-adapters package — new Bedrock embedding models get added over time.
3. Verify the model string isn't accidentally an OpenAI name or a chat model; strip provider prefixes the adapter doesn't expect.

Example fix

// before
await api.embed({ model: 'text-embedding-3-small', input: ['hi'] });

// after
await api.embed({ model: 'bedrock/amazon.titan-embed-text-v2:0', input: ['hi'] });

const SUPPORTED_BEDROCK_EMBED_MODELS = [/^amazon\.titan-embed/, /^cohere\.embed/];
const isSupportedEmbedModel = (m: string) => SUPPORTED_BEDROCK_EMBED_MODELS.some(r => r.test(m));
if (!isSupportedEmbedModel(body.model)) throw new Error(`Unsupported embedding model: ${body.model}`);

try {
  await api.embed(body);
} catch (e) {
  if (e instanceof Error && e.message.startsWith('Unsupported model:')) {
    // switch to a supported embedding model or another provider
  }
  throw e;
}

When it happens

Trigger: Calling embed with a chat model id (e.g. 'anthropic.claude-3-sonnet') or an embedding model id not in the adapter's recognized list; using a newer Bedrock embedding model released after the adapter version in use; passing an OpenAI model name like 'text-embedding-3-small' unmodified.

Common situations: Multi-provider configs reusing OpenAI model names; new Bedrock embedding models not yet supported by the installed adapter version; typos in model identifiers.

Related errors

• Unsupported embeddings type received: number[][]
• Unsupported embeddings type received: number[]
• accessKeyId is required for Bedrock API. Only found…
• AskSage does not support embeddings
• await resp.text()

AI-assisted analysis of continuedev/continue@5522c6f44c (2026-08-27). Data as JSON: /api/errors/e26e2365726aacf3. Report an issue: GitHub.

Appendix: source

Thrown at packages/openai-adapters/src/apis/Bedrock.ts:654

        truncate: "END",
      };
      const output = await this.getInvokeModelResponseBody(body.model, payload);
      embeddings = [output.embedding];
    } else if (body.model.startsWith("amazon.titan-embed")) {
      embeddings = await Promise.all(
        texts.map(async (text) => {
          const payload = {
            inputText: text,
          };
          const output = await this.getInvokeModelResponseBody(
            body.model,
            payload,
          );
          return output.embeddings || [];
        }),
      );
    } else {
      throw new Error(`Unsupported model: ${body.model}`);
    }

    return embedding({
      data: embeddings,
      model: body.model,
      usage: {
        prompt_tokens: 0,
        total_tokens: 0,
      },
    });
  }

  async rerank(body: RerankCreateParams): Promise<CreateRerankResponse> {
    if (!body.query || !body.documents.length) {
      throw new Error("Query and chunks must not be empty");
    }

    // Base payload for both models

View on GitHub (pinned to 5522c6f44c)