chroma-core/chroma · error

Error calling Chroma Embedding API: ${error.message}

Error message

Error calling Chroma Embedding API: ${error.message}

What it means

Top-level error from ChromaCloudQwenEmbeddingFunction.generate(): any Error thrown inside the request block — fetch network failures, response.json() parse failures on non-JSON bodies (HTML error pages), or the inner "Failed to generate embeddings." check — is caught and re-wrapped with the prefix "Error calling Chroma Embedding API:" plus the original message. Read the suffix to find the real cause: network errors mention fetch/ENOTFOUND, parse errors mention JSON, and the bare suffix "Failed to generate embeddings." means the API answered with a JSON body lacking embeddings.

Source

Thrown at clients/new-js/packages/ai-embeddings/chroma-cloud-qwen/src/index.ts:139

      texts,
      instructions: instruction,
    };

    try {
      const response = await fetch(this.url, {
        method: "POST",
        headers: this.headers,
        body: JSON.stringify(snakeCase(body)),
      });

      const data = (await response.json()) as ChromaCloudEmbeddingsResponse;
      if (!data || !data.embeddings) {
        throw new Error("Failed to generate embeddings.");
      }
      return data.embeddings;
    } catch (error) {
      if (error instanceof Error) {
        throw new Error(`Error calling Chroma Embedding API: ${error.message}`);
      } else {
        throw new Error(`Error calling Chroma Embedding API: ${error}`);
      }
    }
  }

  public async generateForQueries(texts: string[]): Promise<number[][]> {
    if (texts.length === 0) {
      return [];
    }

    let instruction = "";
    if (this.task && this.task in this.instructions) {
      instruction =
        this.instructions[this.task][ChromaCloudQwenEmbeddingTarget.QUERY];
    }

    const body: ChromaCloudEmbeddingRequest = {

View on GitHub (pinned to aecdd12c8a)

Solutions

1. Read the suffix after the colon: it names the underlying failure — fix that (network, key, or model).
2. For network causes, verify connectivity to the embed endpoint and proxy settings (HTTPS_PROXY).
3. For "Failed to generate embeddings." suffix, fix the API key / model as per that error.
4. Add retry with backoff around generate() for transient network failures; batch texts so one retry does not re-embed everything.

async function embedWithRetry(ef: { generate(t: string[]): Promise<number[][]> }, texts: string[], tries = 3) {
  for (let i = 1; i <= tries; i++) {
    try {
      return await ef.generate(texts);
    } catch (e) {
      const msg = (e as Error).message;
      const transient = /fetch|network|ENOTFOUND|ECONN|timeout/i.test(msg);
      if (i === tries || !transient) throw e;
      await new Promise((r) => setTimeout(r, 2 ** i * 500));
    }
  }
  throw new Error("unreachable");
}

When it happens

Trigger: fetch() rejects: DNS failure, connection refused, TLS errors, offline; response.json() rejects because a proxy/gateway returned HTML; the inner no-embeddings check fired (auth/model/request errors answered as JSON).

Common situations: Corporate proxies intercepting the embed URL; transient network drops during large batch embedding; misconfigured custom embed URL; expired or missing CHROMA_API_KEY causing JSON error bodies.

Related errors

• Error calling Chroma Embedding API: ${error}
• Error calling Cloudflare Workers AI API: ${error.message}
• Error calling Jina AI API: ${error.message}
• Failed to generate embeddings.
• Error calling Cloudflare Workers AI API: ${error}

AI-assisted analysis of chroma-core/chroma@aecdd12c8a (2026-08-16). Data as JSON: /api/errors/71e8cac4ae380fbf. Report an issue: GitHub.