continuedev/continue · error · Error

Error parsing Ollama response: ${e} ${chunk}

Error message

Error parsing Ollama response: ${e} ${chunk}

What it means

Wrapper error from Ollama._streamComplete's catch block: any failure while parsing or handling a stream chunk — including the inline j.error throw — is rethrown as 'Error parsing Ollama response'. The original error and the offending chunk are embedded in the message, which commonly masks an Ollama-level error as a parse failure.

Source

Thrown at core/llm/llms/Ollama.ts:456

    for await (const value of streamResponse(response)) {
      // Append the received chunk to the buffer
      buffer += value;
      // Split the buffer into individual JSON chunks
      const chunks = buffer.split("\n");
      buffer = chunks.pop() ?? "";

      for (let i = 0; i < chunks.length; i++) {
        const chunk = chunks[i];
        if (chunk.trim() !== "") {
          try {
            const j = JSON.parse(chunk) as OllamaRawResponse;
            if ("error" in j) {
              throw new Error(j.error);
            }
            j.response ??= "";
            yield j.response;
          } catch (e) {
            throw new Error(`Error parsing Ollama response: ${e} ${chunk}`);
          }
        }
      }
    }
  }

  /**
   * Reorder messages so that system messages never appear directly after tool
   * messages. Some Ollama models (Mistral, Ministral) reject the sequence
   * `tool → system` with "Unexpected role 'system' after role 'tool'".
   * This moves such system messages to just before the preceding
   * assistant+tool block.
   */
  private _reorderMessagesForToolCompat(
    messages: OllamaChatMessage[],
  ): OllamaChatMessage[] {
    const result: OllamaChatMessage[] = [...messages];

View on GitHub (pinned to 5522c6f44c)

Solutions

  1. Read the embedded ${e} portion — if it says 'model not found' or similar, fix that underlying issue (ollama pull)
  2. If the chunk looks truncated, check for proxies or timeouts between client and Ollama
  3. Update Ollama to a version whose streaming format matches
Defensive patterns

Strategy: try-catch

Validate before calling

null

Type guard

function unwrapOllamaParseError(e: unknown): string | null {
  const m = e instanceof Error ? e.message.match(/Error parsing Ollama response: (.*)/)?.[1] : null;
  return m ?? null; // inner cause text
}

Try / catch

try { for await (const c of llm.streamComplete(prompt, signal)) yield c; }
catch (e) { const cause = unwrapOllamaParseError(e); if (cause?.includes('not found')) ollamaPull(); else throw e; }

Prevention

When it happens

Trigger: A non-empty stream line that isn't valid JSON, or a valid JSON line containing an error field (which throws inside the try and gets re-wrapped here). Also triggered when the model's output contains malformed NDJSON due to crashes mid-stream.

Common situations: Ollama server killed mid-generation, proxy/load balancer truncating the stream, or the actual root cause being a model-not-found error re-wrapped by this catch.

Related errors

AI-assisted analysis of continuedev/continue@5522c6f44c (2026-08-27). Data as JSON: /api/errors/29cc6fb7ef076872. Report an issue: GitHub.