mem0ai/mem0 · error

LLMReranker requires an LLM instance; RerankerFactory should

Error message

LLMReranker requires an LLM instance; RerankerFactory should always provide one for the llm_reranker provider.

What it means

Thrown by the LLMReranker constructor when it is instantiated without an LLM instance. The llm_reranker provider scores documents by prompting the configured LLM, so RerankerFactory is responsible for passing the memory instance's LLM as the second constructor argument. Seeing this error means the factory/reranker wiring was broken — it is effectively an internal invariant, not a user configuration error.

Source

Thrown at mem0-ts/src/oss/src/rerankers/llm.ts:24

Score the relevance on a scale from 0.0 to 1.0, where:
- 1.0 = Perfectly relevant and directly answers the query
- 0.8-0.9 = Highly relevant with good information
- 0.6-0.7 = Moderately relevant with some useful information
- 0.4-0.5 = Slightly relevant with limited useful information
- 0.0-0.3 = Not relevant or no useful information

Respond with only a single numerical score between 0.0 and 1.0. Do not include any explanation or additional text.`;

const MAX_INPUT_LEN = 4000;

export class LLMReranker implements Reranker {
  private llm: LLM;
  private topK?: number;

  constructor(config: RerankerConfig, llm: LLM) {
    if (!llm) {
      throw new Error(
        "LLMReranker requires an LLM instance; RerankerFactory should always provide one for the llm_reranker provider.",
      );
    }
    this.llm = llm;
    this.topK = config.topK;
  }

  async rerank(
    query: string,
    documents: string[],
    topK?: number,
  ): Promise<RerankResult[]> {
    if (documents.length === 0) return [];

    const scored = await Promise.all(
      documents.map(async (document, index) => {
        try {
          const rerankScore = await this.score(query, document);

View on GitHub (pinned to 001c235229)

Solutions

1. Let RerankerFactory construct the reranker via the Memory config (reranker: { provider: 'llm_reranker', config: { model: ... } }) so the LLM is injected
2. If constructing directly, pass the LLM: new LLMReranker(config, memoryLLM)
3. Update mem0ai to a matching version so factory and reranker signatures agree

Example fix

// before
const reranker = new LLMReranker({ topK: 5 }); // no LLM

// after
const reranker = new LLMReranker({ topK: 5 }, memoryLLM);

if (!llm) throw new Error('llm_reranker requires a configured LLM instance');
const reranker = new LLMReranker(config, llm);

const isLLM = (x: unknown): x is { generateResponse: (...a: unknown[]) => Promise<unknown> } =>
  x != null && typeof (x as { generateResponse?: unknown }).generateResponse === 'function';

When it happens

Trigger: RerankerFactory.create('llm_reranker', config) being called without an LLM (e.g. an older mem0-ts version whose factory does not forward the LLM, or custom code instantiating LLMReranker directly with new LLMReranker(config) and omitting the llm argument).

Common situations: Using a custom or outdated RerankerFactory; writing a custom reranker integration that bypasses the factory; upgrading only part of the monorepo so the factory and reranker signatures disagree.

Related errors

• Unknown LLM provider: ${providerId}
• memory_id is required for update
• AWS Bedrock LLM failed: ${message}
• Azure OpenAI requires both API key and endpoint
• DeepSeek API key is required

AI-assisted analysis of mem0ai/mem0@001c235229 (2026-08-15). Data as JSON: /api/errors/4d4d0020c39169b6. Report an issue: GitHub.