chroma-core/chroma · error · Error

Error calling Jina AI API: ${error.message}

Error message

Error calling Jina AI API: ${error.message}

What it means

The outer catch in JinaEmbeddingFunction.generate() wraps every failure — fetch network errors (DNS, ECONNREFUSED, TLS), JSON parse errors on non-JSON bodies, and even the inner data.detail Error from error 50 (yielding 'Error calling Jina AI API: <server detail>') — into a single prefixed Error. Only the message text survives; the original error class, stack context, and cause chain are lost.

Source

Thrown at clients/js/packages/chromadb-core/src/embeddings/JinaEmbeddingFunction.ts:132

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

      const data = (await response.json()) as { data: any[]; detail: string };
      if (!data || !data.data) {
        throw new Error(data.detail);
      }

      const embeddings: any[] = data.data;
      const sortedEmbeddings = embeddings.sort((a, b) => a.index - b.index);

      return sortedEmbeddings.map((result) => result.embedding);
    } catch (error) {
      if (error instanceof Error) {
        throw new Error(`Error calling Jina AI API: ${error.message}`);
      } else {
        throw new Error(`Error calling Jina AI API: ${error}`);
      }
    }
  }

  buildFromConfig(config: StoredConfig): JinaEmbeddingFunction {
    return new JinaEmbeddingFunction({
      model_name: config.model_name,
      api_key_env_var: config.api_key_env_var,
      task: config.task,
      late_chunking: config.late_chunking,
      truncate: config.truncate,
      dimensions: config.dimensions,
      embedding_type: config.embedding_type,
      normalized: config.normalized,
    });
  }

View on GitHub (pinned to aecdd12c8a)

Solutions

1. Inspect the text after 'Error calling Jina AI API:' — it carries the root cause (e.g. 'fetch failed', 'Invalid API key')
2. Reproduce connectivity from the same host/container: curl -sS https://api.jina.ai/v1/embeddings to distinguish network from auth issues
3. Add api.jina.ai to proxy/egress allowlists or fix proxy env vars
4. If the suffix is a Jina detail string, apply the fix for that underlying error (key, quota, model)

Example fix

// before (library code): opaque wrapper loses the cause
throw new Error(`Error calling Jina AI API: ${error.message}`);

// after (library-level): rethrow untouched so stack/cause survive
throw error;

async function jinaReachable(): Promise<boolean> {
  try {
    await fetch("https://api.jina.ai", { method: "HEAD", signal: AbortSignal.timeout(5_000) });
    return true; // any HTTP answer proves DNS/TLS/egress work
  } catch {
    return false;
  }
}
// run at startup in restricted-network environments

try {
  const vecs = await ef.generate(texts);
} catch (e) {
  const msg = e instanceof Error ? e.message : String(e);
  if (msg.startsWith("Error calling Jina AI API:")) {
    const cause = msg.slice("Error calling Jina AI API:".length).trim();
    // classify: 'fetch failed' -> network/egress; anything else -> server detail (auth/quota/model)
  }
  throw e;
}

When it happens

Trigger: api.jina.ai unreachable (offline environment, blocked egress, proxy misconfiguration); a proxy/captive portal returning an HTML error page so response.json() throws; or the inner !data.data branch throwing first and being re-wrapped here.

Common situations: Corporate proxies blocking api.jina.ai; air-gapped environments; misconfigured HTTPS_PROXY; double-wrapped detail errors making logs noisy and hard to alert on.

Related errors

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

AI-assisted analysis of chroma-core/chroma@aecdd12c8a (2026-08-16). Data as JSON: /api/errors/7a95397a219f0964. Report an issue: GitHub.