mem0ai/mem0 · error · Error

AWS Bedrock model ${this.model} returned no embedding for on

Error message

AWS Bedrock model ${this.model} returned no embedding for one or more inputs

What it means

Thrown by the AWS Bedrock embedder after a successful invoke when the response shape is unusable: no embeddings array at all, a count that does not match the number of input texts, or at least one zero-length vector. The explicit zero-length check exists because an empty array is truthy in JS and would otherwise slip through a naive length check and hand the caller a degenerate embedding.

Source

Thrown at mem0-ts/src/oss/src/embeddings/aws_bedrock.ts:255

    // Validated outside the try so this message is not re-wrapped by the catch.
    // Cohere v3 replies with a flat `embeddings` array; v4 (when
    // embedding_types is requested) nests it under `.float`.
    const embeddings = this.isCohereModel()
      ? Array.isArray(payload.embeddings)
        ? payload.embeddings
        : payload.embeddings?.float
      : payload.embedding && [payload.embedding];

    // `[]` is truthy, so a lone zero-length vector must be checked for
    // explicitly -- otherwise it passes the length check and hands the
    // caller an empty embedding instead of an error.
    if (
      !embeddings ||
      embeddings.length !== texts.length ||
      embeddings.some((embedding) => embedding.length === 0)
    ) {
      throw new Error(
        `AWS Bedrock model ${this.model} returned no embedding for one or more inputs`,
      );
    }
    return embeddings;
  }

  async embed(
    text: string,
    memoryAction?: "add" | "update" | "search",
  ): Promise<number[]> {
    return (await this.invoke([text], memoryAction))[0];
  }

  async embedBatch(
    texts: string[],
    memoryAction?: "add" | "update" | "search",
  ): Promise<number[][]> {
    if (texts.length === 0) return [];

View on GitHub (pinned to 001c235229)

Solutions

1. Log texts.length and the raw payload (or catch and inspect) to see which of the three conditions fired: missing / count mismatch / zero-length vector
2. Filter empty or whitespace-only strings from the batch before embedding
3. Reduce batch size below the model's per-request input limit (Cohere embed on Bedrock allows far fewer than unbounded arrays)
4. Confirm the configured model name matches the response format you expect — do not mix a v4 model ID with v3 handling assumptions

Example fix

// before
const vectors = await embedder.embedBatch(allTexts); // 1 empty string in batch -> zero-length vector

// after
const clean = allTexts.map((t) => t.trim()).filter((t) => t.length > 0);
const vectors = await embedder.embedBatch(clean);

const clean = texts.map((t) => t.trim()).filter((t) => t.length > 0);
if (clean.length !== texts.length) {
  // decide how to map vectors back if you drop inputs
  console.warn(`dropped ${texts.length - clean.length} empty inputs`);
}
await embedder.embedBatch(clean);

try {
  vectors = await embedder.embedBatch(batch);
} catch (e) {
  if (e instanceof Error && e.message.includes('returned no embedding')) {
    // data-integrity failure: shrink the batch and retry once; do not use partial vectors
    vectors = await embedder.embedBatch(batch.slice(0, Math.ceil(batch.length / 2)));
  } else throw e;
}

When it happens

Trigger: Bedrock returns partial results (batch throttled mid-response); a Cohere v3 flat payload vs v4 payload-shape mismatch; payload.embeddings undefined because the response body was an error object that still parsed as JSON; one input string rejected (e.g. empty text after preprocessing) yielding a null/empty vector.

Common situations: Sending a batch larger than the model limit so the service returns fewer vectors; empty-string inputs in the batch; model family whose response format changed (Cohere v3 'embeddings' vs v4 nested '.float'); intermittent partial responses under load.

Related errors

• Error getting embedding from AWS Bedrock model ${this.model}
• Error getting embedding from AWS Bedrock: {e}
• AWS Bedrock requires both awsAccessKeyId and awsSecretAccess
• The 'boto3' library is required. Please install it using 'pi
• Failed to generate response: {e}

AI-assisted analysis of mem0ai/mem0@001c235229 (2026-08-15). Data as JSON: /api/errors/ff5547072fad72a9. Report an issue: GitHub.