googleapis/mcp-toolbox · error

error embedding parameters with model %s: %w

Error message

error embedding parameters with model %s: %w

What it means

EmbedParams wraps any error returned by the embedding model's EmbedParameters API call with this message, preserving the root cause via %w. The model itself resolved fine, but the embedding request failed — the real cause (network, auth, provider error) is in the wrapped error after this prefix.

Source

Thrown at internal/util/parameters/parameters.go:220

		})
	}

	// Batch embedding request sent to each model
	for modelName, params := range parametersToEmbed {
		model, ok := pMgr.GetEmbeddingModel(modelName)
		if !ok {
			return nil, fmt.Errorf("embedding model does not exist: %s", modelName)
		}

		// Extract only the string values for the API call
		stringBatch := make([]string, len(params))
		for i, paramStr := range params {
			stringBatch[i] = paramStr.OriginalValue
		}

		embeddings, err := model.EmbedParameters(ctx, stringBatch)
		if err != nil {
			return nil, fmt.Errorf("error embedding parameters with model %s: %w", modelName, err)
		}

		if len(embeddings) != len(stringBatch) {
			return nil, fmt.Errorf("model %s returned %d embeddings for %d inputs", modelName, len(embeddings), len(stringBatch))
		}

		for i, rawVector := range embeddings {

			item := params[i]

			// Call vector formatter
			var finalValue any = rawVector

			if formatter == nil {
				paramValues[item.Index].Value = finalValue
				continue
			}

View on GitHub (pinned to 8cc6e09de2)

Solutions

1. Unwrap the error (errors.Unwrap / %w chain) and read the underlying provider message; fix that first.
2. Verify provider credentials (API key env vars, Application Default Credentials) are present and valid.
3. Test network connectivity to the embedding provider endpoint from the runtime environment.
4. Reduce batch size and add exponential backoff for rate-limit (429) responses; check quotas.
5. Confirm the model kind/name in config is one the provider actually supports.

Example fix

// before: opaque wrapper only
if err != nil { return err }

// after: surface the root cause
embeddings, err := model.EmbedParameters(ctx, batch)
if err != nil {
    return fmt.Errorf("embedding failed: %w", err) // inspect wrapped cause
}

// pre-flight at startup
if _, err := model.EmbedParameters(ctx, []string{"ping"}); err != nil {
    log.Fatalf("embedding provider unavailable: %v", err)
}

embeddings, err := model.EmbedParameters(ctx, batch)
if err != nil {
    if isRetryable(err) { // 429/5xx
        time.Sleep(backoff); return retry(batch)
    }
    return fmt.Errorf("embedding unavailable: %w", err)
}

When it happens

Trigger: model.EmbedParameters(ctx, stringBatch) returns a non-nil error during a batch embedding call: network failure to the provider, missing/invalid API credentials, quota or rate-limit rejection, or a provider 4xx/5xx response.

Common situations: Expired or absent provider API keys (env vars / ADC not set in the deployment environment); blocked egress in VPC or Cloud Run; hitting provider rate limits with large parameter batches; misconfigured model endpoint or project settings.

Related errors

• failed to verify allowedDataset '%s' in project '%s': %w
• error fetching connections: %w
• error fetching LookML models: %w
• error embedding parameters: %w
• toolbox failed to start listener: %w

AI-assisted analysis of googleapis/mcp-toolbox@8cc6e09de2 (2026-09-05). Data as JSON: /api/errors/9096eac8dfa02f75. Report an issue: GitHub.