Enabling Dynamic Message Insertion: Extend Chat Memory Manager to Accept JSON Arrays
The Chat Memory Manager node's inability to accept dynamic JSON arrays for inserting messages hampers integrations with external chat stores and variable-length histories. Supporting array inputs would simplify RAG workflows, chat migrations, and programmatic control of memory with fewer brittle UI workarounds.
The current UI-centric approach - requiring a fixed number of message rows - creates friction when chat histories arrive as arrays from APIs or databases, or when the number of messages varies per execution. This limitation forces developers to hardcode indices or build iterative loops in separate function nodes, adding complexity, runtime overhead, and potential for errors.
From a product perspective, the fix is straightforward: accept a dynamic JSON array as an input parameter, map array elements to the internal message schema (type, role, content, metadata), and validate each element against a schema. Feature considerations should include batch insertion with size limits, optional deduplication, and streaming inserts for very large histories. Exposing a mapping/transformation option for custom fields (timestamps, source flags) would increase flexibility for diverse sources.
Short-term workarounds include programmatically transforming the array into the node's expected format using loop or function nodes, or creating a small custom node that performs the mapping and invokes the Chat Memory Manager endpoint. But these add maintenance burden and complexity for non-engineering users.
For business leaders and product owners: prioritize this enhancement if you operate RAG assistants, migrate chat histories, or integrate third-party conversational logs. It reduces development time, lowers error surface, and improves the UX for integrating heterogeneous chat sources - all of which accelerate time-to-value for conversational AI initiatives.
Original Source
n8n Community