What causes "Memory with ID ${memoryId} not found" in mem0ai/mem0?

Verify the ID exists first with memory.get(memoryId) and skip gracefully if null

Memory with ID ${memoryId} not found

Error message

Memory with ID ${memoryId} not found

What it means

Internal guard in the private updateMemory(): the memory ID is not present in the vector store (vectorStore.get(id) returned null) before attempting an update. It usually surfaces through the public update()/add() reconciliation flow when a stale or already-deleted memory ID is referenced, e.g. in linked_memory_ids or an inferred-update path.

Source

Thrown at mem0-ts/src/oss/src/memory/index.ts:1959

      memoryId,
      null,
      data,
      "ADD",
      memoryMetadata.createdAt,
    );

    return memoryId;
  }

  private async updateMemory(
    memoryId: string,
    data: string | undefined,
    existingEmbeddings: Record<string, number[]>,
    metadata: Record<string, any> = {},
  ): Promise<string> {
    const existingMemory = await this.vectorStore.get(memoryId);
    if (!existingMemory) {
      throw new Error(`Memory with ID ${memoryId} not found`);
    }

    const prevValue = existingMemory.payload.data;
    // Metadata-only update: fall back to the stored text so we can re-index it.
    const newData = data ?? prevValue;
    if (typeof newData !== "string") {
      throw new Error(
        `Memory with ID ${memoryId} does not have text content to update`,
      );
    }
    const textChanged = newData !== prevValue;

    const embedding = Object.prototype.hasOwnProperty.call(
      existingEmbeddings,
      newData,
    )
      ? existingEmbeddings[newData]
      : await this.embedder.embed(newData, "update");

View on GitHub (pinned to 001c235229)

Solutions

Verify the ID exists first with memory.get(memoryId) and skip gracefully if null
If IDs are persisted long-term, re-fetch them per session instead of caching
In multi-writer setups, coordinate resets so no client holds stale IDs

Example fix

// before
await memory.update(staleId, { text: 'updated' });

// after
const mem = await memory.get(staleId);
if (mem) {
  await memory.update(staleId, { text: 'updated' });
}

Defensive patterns

Strategy: try-catch

Validate before calling

const existing = await memory.get(memoryId);
if (!existing) {
  // stale reference — refresh your memory list and skip
}

Try / catch

try { await memory.update(id, { text }); } catch (e) { if (e instanceof Error && e.message.includes('not found')) { await refreshMemoryList(); return; } throw e; }

Prevention

Do not cache memory IDs across store resets or long sessions — re-fetch before mutating
In multi-instance deployments, coordinate who may call reset() while others hold references

When it happens

Trigger: updateMemory(memoryId) is called with an ID that does not exist in the vector store — deleted by another process, expired and cleaned up, from a different environment, or a typo'd/hallucinated ID (LLM extraction can emit IDs that no longer resolve).

Common situations: Holding a memory ID across a long session while another client deletes or resets the store; multi-instance deployments where one instance resets the vector DB; LLM extraction returning linked IDs from a previous conversation state.

Related errors

AI-assisted analysis of mem0ai/mem0@001c235229 (2026-08-15). Data as JSON: /api/errors/9a6847564d0ea1b1. Report an issue: GitHub.