vercel/ai · error · TooManyEmbeddingValuesForCallError

Too many values for a single embedding call. The ${provider}

Error message

Too many values for a single embedding call. The ${provider} model "${modelId}" can only embed up to ${maxEmbeddingsPerCall} values per call, but ${values.length} values were provided.

What it means

The Voyage embedding model enforces a hard limit (maxEmbeddingsPerCall = 128) on how many values can be embedded in a single API call; exceeding it throws TooManyEmbeddingValuesForCallError before any request is sent. Higher-level embedMany batch settings must respect this per-call cap.

Source

Thrown at packages/voyage/src/voyage-embedding-model.ts:76

    return this.config.provider;
  }

  async doEmbed({
    values,
    headers,
    abortSignal,
    providerOptions,
  }: Parameters<EmbeddingModelV4['doEmbed']>[0]): Promise<
    Awaited<ReturnType<EmbeddingModelV4['doEmbed']>>
  > {
    const embeddingOptions = await parseProviderOptions({
      provider: 'voyage',
      providerOptions,
      schema: voyageEmbeddingModelOptions,
    });

    if (values.length > this.maxEmbeddingsPerCall) {
      throw new TooManyEmbeddingValuesForCallError({
        provider: this.provider,
        modelId: this.modelId,
        maxEmbeddingsPerCall: this.maxEmbeddingsPerCall,
        values,
      });
    }

    const {
      responseHeaders,
      value: response,
      rawValue,
    } = await postJsonToApi({
      url: `${this.config.baseURL}/embeddings`,
      headers: combineHeaders(this.config.headers?.(), headers),
      body: {
        input: values,
        model: this.modelId,
        input_type: embeddingOptions?.inputType,

View on GitHub (pinned to 69428b1f8b)

Solutions

Split values into chunks of at most 128 and call embed per chunk (or set embedMany's maxEmbeddingsPerCall accordingly).
Set maxEmbeddingsPerCall: 128 in the embedMany settings so batching is automatic.
Add validation of values.length before calling embed.
Compare limits when migrating providers and adjust batch sizes per provider.

Example fix

// before
await embedMany({ model: voyage.textEmbeddingModel('voyage-3'), values: bigArray });
// after
await embedMany({
  model: voyage.textEmbeddingModel('voyage-3'),
  values: bigArray,
  maxEmbeddingsPerCall: 128,
});

Defensive patterns

Strategy: validation

Validate before calling

const MAX = 128;
if (values.length > MAX) {
  throw new Error(`Voyage embeds at most ${MAX} values per call; got ${values.length}.`);
}

Type guard

null

Try / catch

try {
  await embed({ model, values });
} catch (error) {
  if (TooManyEmbeddingValuesForCallError.isInstance(error)) {
    // chunk values and retry
  } else throw error;
}

Prevention

Chunk large value arrays into <=128 before embedding.
Set maxEmbeddingsPerCall in embedMany so batching is automatic.
Check per-call limits when switching embedding providers.
Add startup tests embedding a 129-item batch in CI.

When it happens

Trigger: Calling embed()/embedMany() with more than 128 values for a single call against a Voyage embedding model (or a maxEmbeddingsPerCall configured below the provided count).

Common situations: Bulk embedding of a large document set in one call; migrating from a provider with a higher per-call limit; incorrectly configured batch size in embedMany options.

Related errors

AI-assisted analysis of vercel/ai@69428b1f8b (2026-08-30). Data as JSON: /api/errors/8366325ee0dd2f24. Report an issue: GitHub.