Mintplex-Labs/anything-llm · error · Error

Ollama Failed to embed: ${error}

Error message

Ollama Failed to embed: ${error}

What it means

Thrown at the end of embedChunks (line 133) if the per-batch loop set `error` — which happens in the catch on any batch failure, including the empty-embeddings throw. The loop breaks on first error, clears data, and re-throws a single aggregated 'Ollama Failed to embed: <message>' so the caller sees one failure rather than partial results.

Source

Thrown at server/utils/EmbeddingEngines/ollama/index.js:133

          throw new Error("Ollama returned empty embeddings for batch!");

        // Using prompt param in embed() would return a single embedding (number[])
        // but input param returns an array of embeddings (number[][]) for batch processing.
        // This is why we spread the embeddings array into the data array.
        data.push(...embeddings);
        reportEmbeddingProgress(data.length, textChunks.length);
        this.log(
          `Batch ${currentBatch}/${totalBatches}: Embedded ${embeddings.length} chunks. Total: ${data.length}/${textChunks.length}`
        );
      } catch (err) {
        this.log(err.message);
        error = err.message;
        data = [];
        break;
      }
    }

    if (!!error) throw new Error(`Ollama Failed to embed: ${error}`);
    return data.length > 0 ? data : null;
  }
}

module.exports = {
  OllamaEmbedder,
};

View on GitHub (pinned to 526360e320)

Solutions

1. Read the embedded message: 'empty embeddings for batch' -> switch to a real embedding model (#335); 'model not found' -> `ollama pull <model>`; connection errors -> see #334
2. Increase num_ctx by raising the configured max chunk length if chunks are being truncated
3. Free Ollama resources (close other models) to avoid unload/reload thrash on low-RAM hosts
4. For a remote proxy, verify OLLAMA_AUTH_TOKEN and network stability

Example fix

// before
// low-RAM host, model thrashing between batches

// after
// pin one embedding model and free others
ollama stop <chat-model>
EMBEDDING_MODEL_PREF=nomic-embed-text

// reserve enough context so chunks aren't truncated to nothing
function saneChunkLength(n) { return Number.isFinite(n) && n > 0; }
if (!saneChunkLength(maximumChunkLength())) {
  throw new Error('embeddingMaxChunkLength is non-positive; chunks would be truncated.');
}

function isOllamaEmbedError(e) {
  return e instanceof Error && /Ollama Failed to embed/.test(e.message);
}

try {
  return await embedder.embedChunks(chunks);
} catch (e) {
  if (/model not found|empty embeddings/.test(e.message)) throw e;  // fix config
  // resource/connection errors may clear on retry
  await new Promise(r => setTimeout(r, 2000));
  return await embedder.embedChunks(chunks);
}

When it happens

Trigger: Any batch's try block throws: client.embed network rejection; the empty-embeddings defensive throw (#335); model not found; Ollama daemon crashed mid-loop; context window exceeded (num_ctx too small for a chunk); auth token rejected for a remote Ollama.

Common situations: Ollama OOM mid-batch on low-memory machines (the code comment explicitly calls this out); model unloaded between batches; embeddingMaxChunkLength too small causing truncation errors; remote Ollama connection flapping; OLLAMA_AUTH_TOKEN wrong for a proxied instance.

Related errors

• LiteLLM Failed to embed: ${error}
• LMStudio Failed to embed: ${Array.from(uniqueErrors).join(",
• LocalAI Failed to embed: ${error}
• Ollama service could not be reached. Is Ollama running?
• Ollama returned empty embeddings for batch!

AI-assisted analysis of Mintplex-Labs/anything-llm@526360e320 (2026-08-13). Data as JSON: /api/errors/3c53ae46507b1823. Report an issue: GitHub.