mem0ai/mem0 · error · Error

DeepSeek LLM failed: ${message}

Error message

DeepSeek LLM failed: ${message}

What it means

Thrown by DeepSeekLLM.generateResponse when the inherited OpenAILLM call to DeepSeek's OpenAI-compatible API fails for any reason. The wrapper preserves the original error text after the 'DeepSeek LLM failed:' prefix, so the suffix identifies whether it was auth (401), rate limit (429), model issues, JSON mode problems, or network.

Source

Thrown at mem0-ts/src/oss/src/llms/deepseek.ts:31

      apiKey,
      baseURL:
        config.baseURL ||
        process.env.DEEPSEEK_API_BASE ||
        "https://api.deepseek.com",
      model: config.model || "deepseek-chat",
    });
  }

  async generateResponse(
    messages: Message[],
    responseFormat?: { type: string },
    tools?: any[],
  ): Promise<string | LLMResponse> {
    try {
      return await super.generateResponse(messages, responseFormat, tools);
    } catch (err) {
      const message = err instanceof Error ? err.message : String(err);
      throw new Error(`DeepSeek LLM failed: ${message}`);
    }
  }

  async generateChat(messages: Message[]): Promise<LLMResponse> {
    try {
      return await super.generateChat(messages);
    } catch (err) {
      const message = err instanceof Error ? err.message : String(err);
      throw new Error(`DeepSeek LLM failed: ${message}`);
    }
  }
}

View on GitHub (pinned to 001c235229)

Solutions

1. Inspect the message after the prefix: 401 → fix the key, 402 → top up balance, 429 → back off, model errors → correct the model string.
2. Verify the key works outside mem0: curl https://api.deepseek.com/models with Authorization: Bearer $DEEPSEEK_API_KEY.
3. For 429s, wrap calls in exponential-backoff retry (DeepSeek rate limits are per-minute TPM/RPM).
4. If using a custom baseURL, confirm the proxy is reachable and OpenAI-compatible from the runtime environment.
5. Confirm the model value: defaults to deepseek-chat; deepseek-reasoner rejects some parameters (e.g. temperature).

Example fix

// before
const res = await deepSeekLlm.generateResponse(messages, { type: 'json_object' });
// DeepSeek LLM failed: 402 Insufficient Balance

// after (fail fast on balance/auth, retry on transient 429)
try {
  const res = await deepSeekLlm.generateResponse(messages, { type: 'json_object' });
} catch (e) {
  const msg = e instanceof Error ? e.message : String(e);
  if (msg.includes('402') || msg.includes('401')) throw e; // fatal: billing/auth
  if (msg.includes('429')) await backoffThenRetry(() => deepSeekLlm.generateResponse(messages));
  else throw e;
}

async function deepSeekReachable(base = 'https://api.deepseek.com') {
  const r = await fetch(`${base}/models`, { headers: { Authorization: `Bearer ${process.env.DEEPSEEK_API_KEY}` } });
  if (r.status === 401) throw new Error('DeepSeek key invalid');
  if (r.status === 402) throw new Error('DeepSeek balance exhausted');
  return r.ok;
}

function isDeepSeekFatal(err: unknown): boolean {
  const m = err instanceof Error ? err.message : String(err);
  return m.startsWith('DeepSeek LLM failed:') && /401|402|invalid/i.test(m);
}

try {
  return await deepSeekLlm.generateResponse(messages, responseFormat);
} catch (err) {
  const msg = String(err);
  if (/429|timeout|ECONN/i.test(msg)) return retry(fn, { retries: 3 });
  throw err; // 401/402/model errors are fatal
}

When it happens

Trigger: Calling generateResponse() on DeepSeekLLM (directly or via Memory.add/search/ update flows) when DeepSeek returns 401 invalid key, 402 insufficient balance, 429 rate limit, an invalid model name, a malformed responseFormat for JSON mode, or when the network call to https://api.deepseek.com fails or times out.

Common situations: Expired or revoked DeepSeek key; DeepSeek account out of credits (402 is common on the free-trial exhaustion); switching model between deepseek-chat and deepseek-reasoner with parameters only one supports; concurrent requests exceeding the per-key rate limit; custom DEEPSEEK_API_BASE pointing at a proxy that returns HTML errors.

Related errors

• DeepSeek API key is required
• LiteLLM failed: ${message}
• LM Studio LLM failed: ${message}
• MiniMax LLM failed: ${message}
• Sarvam LLM failed: ${message}

AI-assisted analysis of mem0ai/mem0@001c235229 (2026-08-15). Data as JSON: /api/errors/833414bf94aa5bfa. Report an issue: GitHub.