n8n 2.47.6 fixes Redis caching, Pinecone search, and expression speed

n8n released version 2.47.6 on November 13, 2026. The patch combines bug fixes, one feature, and performance work across the core engine, the editor, and the Pinecone Vector Store node.

On the core side, trusted-source cache keys now stay in a single Redis cluster slot. AI agents can also turn off prompt caching for endpoints that reject the cache fields, and legacy-format data is read with the instance key when the stored key differs. In the editor, JSON highlight color is now visible on the light theme. The Gateway config can now show Gateway credits promotions.

The Pinecone node can now search indexes that use a document schema through the documents API. On the performance side, the release moves the replaceSpecialChars character map out of the runtime bundle and reduces per-item setup in the expression runtime.

Why it matters

Patch releases often fix edge cases that cause failed or unexpected workflow runs. This release targets caching, vector search, and expression evaluation, all of which can affect production automations that rely on those features.

What to do

If you use Redis cluster caching, Pinecone document indexes, or AI agents with prompt caching, review the release notes before upgrading. The notes do not describe migration steps, so confirm upgrade guidance on the vendor’s site and test key workflows in a staging environment first.