What causes "e.message" in Mintplex-Labs/anything-llm?

Match the error text to a status: 401 -> fix the key, 429 -> back off, 404 -> repick the model, network -> check egress

e.message

Error message

e.message

What it means

Not a fixed message: this is the .catch attached to openai.chat.completions.create re-throwing the upstream ApiPie (OpenAI-compatible) error's message verbatim. Whatever the endpoint returned - 401 incorrect APIPIE_LLM_API_KEY, 429 quota, 404 model not found, socket timeouts, oversized prompts - becomes the Error text. Because the promise rejects before completion, LLMPerformanceMonitor never records metrics for the call.

Solutions

Match the error text to a status: 401 -> fix the key, 429 -> back off, 404 -> repick the model, network -> check egress
Verify APIPIE_LLM_API_KEY with a direct curl to https://apipie.ai/v1/models
Retry with exponential backoff on 429/5xx
Trim workspace context if context-length errors appear

Example fix

// before
const result = await apipie.getChatCompletion(messages);

// after
try {
  const result = await apipie.getChatCompletion(messages);
} catch (e) {
  if (/429|rate/i.test(e.message)) {
    await new Promise((r) => setTimeout(r, 5000));
    return apipie.getChatCompletion(messages);
  }
  throw e;
}

Defensive patterns

Strategy: try-catch

Validate before calling

// Cheap liveness + auth probe before a long chat call
try {
  await apipie.openai.models.list();
} catch (e) {
  return res.status(502).json({ error: `ApiPie unreachable or key invalid: ${e.message}` });
}

Try / catch

try {
  const result = await apipie.getChatCompletion(messages);
} catch (e) {
  if (/401|api key/i.test(e.message)) handleBadKey();
  else if (/429|quota|rate/i.test(e.message)) return retryWithBackoff(fn);
  else if (/404|model/i.test(e.message)) handleBadModel();
  else throw e;
}

Prevention

Probe the key with models.list before starting long conversations
Classify upstream errors by inspecting the message text (it is verbatim from the API)
Bound prompt size to avoid context-length rejections
Retry transient statuses only, with backoff

When it happens

Trigger: Invalid or expired ApiPie key (401); model removed between validation and the call (404); rate/quota limits (429); network failure reaching apipie.ai; prompt exceeding the upstream model's context window.

Common situations: Key rotation without updating provider settings; aggregator or upstream provider outages; long workspace context pushed into the prompt; regional or firewall blocks on apipie.ai.

Related errors

AI-assisted analysis of Mintplex-Labs/anything-llm@3aec848f28 (2026-08-18). Data as JSON: /api/errors/82f5c2a0b30a0058. Report an issue: GitHub.

Appendix: source

Thrown at server/utils/AiProviders/apipie/index.js:204

      },
    ];
  }

  async getChatCompletion(messages = null, { temperature = 0.7 }) {
    if (!(await this.isValidChatCompletionModel(this.model)))
      throw new Error(
        `ApiPie chat: ${this.model} is not valid for chat completion!`
      );

    const result = await LLMPerformanceMonitor.measureAsyncFunction(
      this.openai.chat.completions
        .create({
          model: this.model,
          messages,
          temperature,
        })
        .catch((e) => {
          throw new Error(e.message);
        })
    );

    if (
      !result.output.hasOwnProperty("choices") ||
      result.output.choices.length === 0
    )
      return null;

    return {
      textResponse: result.output.choices[0].message.content,
      metrics: {
        prompt_tokens: result.output.usage?.prompt_tokens || 0,
        completion_tokens: result.output.usage?.completion_tokens || 0,
        total_tokens: result.output.usage?.total_tokens || 0,
        outputTps:
          (result.output.usage?.completion_tokens || 0) / result.duration,
        duration: result.duration,

View on GitHub (pinned to 3aec848f28)