How do I import "Adaptive RAG" into n8n?

Click Download n8n JSON on this page, then in n8n choose Workflows → Import from file and select the downloaded file. Reconnect credentials on each node — the export carries credential IDs from the original author's instance, not working secrets.

Which integrations does "Adaptive RAG" use?

It uses: agent, switch, chat, lmChatGoogleGemini, memoryBufferWindow, embeddingsGoogleGemini, summarize, vectorStoreQdrant. The flow diagram above shows how each service connects across the 39 nodes. Filter OpenRuna by integration to find related workflows and tools.

Can I import a sanitized export without credentials?

Yes. Use "Download sanitized JSON" to strip credential references before sharing internally. You will still add your own agent and switch credentials in n8n after import.

What trigger type does "Adaptive RAG" use?

Trigger type: scheduled/triggered. After import, confirm the trigger node (webhook, schedule, or app event) matches your environment, then activate the workflow when you are ready to run it.

Is "Adaptive RAG" free to use?

OpenRuna catalogs community n8n workflows under open licenses. Check the source corpus link for the original license; commercial use may require verifying the third-party API terms for agent, switch, chat, lmChatGoogleGemini, memoryBufferWindow, embeddingsGoogleGemini, summarize, vectorStoreQdrant.

← Workflows

Adaptive RAG

39 nodes · agent, switch, chat, lmChatGoogleGemini, memoryBufferWindow, embeddingsGoogleGemini

Engine: n8n
Nodes: 39
Trigger: scheduled/triggered
Source file: 1829_Summarize_Respondtowebhook_Automation_Webhook.json

Workflow flow

IntegrationsAgentSwitchChatLmChatGoogleGeminiMemoryBufferWindowEmbeddingsGoogleGeminiSummarizeVectorStoreQdrantRespondToWebhookExecuteWorkflow@n8n/n8n-nodes-langchain.agent@n8n/n8n-nodes-langchain.chat+5 more

Integration path in execution order (sticky notes and duplicate steps omitted)

@n8n/n8n-nodes-langchain.agentSwitch@n8n/n8n-nodes-langchain.chatSet@n8n/n8n-nodes-langchain.lmChatGoogleGemini@n8n/n8n-nodes-langchain.memoryBufferWindow@n8n/n8n-nodes-langchain.embeddingsGoogleGeminiSummarize@n8n/n8n-nodes-langchain.vectorStoreQdrantRespondToWebhookExecuteWorkflow

30 steps · use Full diagram for every node

Browse by integration

agent switch chat lmChatGoogleGemini memoryBufferWindow embeddingsGoogleGemini summarize vectorStoreQdrant respondToWebhook executeWorkflow

Browse all integrations →

Download n8n JSON Download sanitized JSON Download flow graph

Import into n8n

Download the n8n JSON file (full or sanitized).
In n8n: Workflows → Import from file.
Open each node and reconnect credentials to your accounts.
Test with manual execution before enabling triggers.

Sanitized JSON removes credential IDs — safer for sharing. Flow graph JSON is OpenRuna's lightweight format for diagrams and other tools, not n8n import.

View source corpus ↗

Overview

Adaptive RAG is a free n8n workflow on OpenRuna — 39 nodes connecting agent, switch, chat, lmChatGoogleGemini, memoryBufferWindow, embeddingsGoogleGemini, summarize, vectorStoreQdrant. Download the JSON, view the flow diagram, and import it into your n8n instance.

What this workflow does

"Adaptive RAG" automates a agent-to-switch flow in n8n across 39 nodes. 39 nodes · agent, switch, chat, lmChatGoogleGemini, memoryBufferWindow, embeddingsGoogleGemini It starts from a scheduled/triggered trigger and routes data through every node in order. On OpenRuna you get the visual graph, the integration list, and a downloadable JSON ready to import into n8n. Use it to learn integration patterns, fork it for your own use case, or deploy it quickly after reconnecting credentials.

Use cases

Benchmark it against your current agent integration to spot missing steps or redundant nodes.
Fork it to add error handling, retries, or a notification step without designing the 39-node flow from scratch.
Share it with your team as a documented reference for scheduled/triggered automations in the agent stack.
Study how data moves from agent to switch before building your own variant.

Example output

Imported into n8n, "Adaptive RAG" exposes 39 configured nodes. Trigger type: scheduled/triggered. Integrations: agent, switch, chat, lmChatGoogleGemini, memoryBufferWindow, embeddingsGoogleGemini, summarize, vectorStoreQdrant. Reconnect the agent credential first, then walk the remaining nodes top to bottom. Run it once in test mode to verify any branch logic (Switch/IF nodes) before enabling production triggers.

Tips by platform

Claude

Paste the exported JSON for "Adaptive RAG" into Claude and ask it to explain each of the 39 nodes. Claude is good at suggesting credential mappings for agent, switch, chat, lmChatGoogleGemini, memoryBufferWindow, embeddingsGoogleGemini, summarize, vectorStoreQdrant, plus error-handling and idempotency improvements before you run it in production.

ChatGPT

Have ChatGPT generate a checklist for adapting "Adaptive RAG" to your accounts — which agent, switch, chat, lmChatGoogleGemini, memoryBufferWindow, embeddingsGoogleGemini, summarize, vectorStoreQdrant credentials to create and which node parameters to change first.

Cursor

Open the downloaded JSON in Cursor and use Agent mode to refactor hard-coded URLs into env vars, or to generate tests for any Code nodes. Keep the OpenRuna flow diagram open so you do not break connection order across the 39 nodes.

Frequently asked questions

How do I import "Adaptive RAG" into n8n?: Click Download n8n JSON on this page, then in n8n choose Workflows → Import from file and select the downloaded file. Reconnect credentials on each node — the export carries credential IDs from the original author's instance, not working secrets.
Which integrations does "Adaptive RAG" use?: It uses: agent, switch, chat, lmChatGoogleGemini, memoryBufferWindow, embeddingsGoogleGemini, summarize, vectorStoreQdrant. The flow diagram above shows how each service connects across the 39 nodes. Filter OpenRuna by integration to find related workflows and tools.
Can I import a sanitized export without credentials?: Yes. Use "Download sanitized JSON" to strip credential references before sharing internally. You will still add your own agent and switch credentials in n8n after import.
What trigger type does "Adaptive RAG" use?: Trigger type: scheduled/triggered. After import, confirm the trigger node (webhook, schedule, or app event) matches your environment, then activate the workflow when you are ready to run it.
Is "Adaptive RAG" free to use?: OpenRuna catalogs community n8n workflows under open licenses. Check the source corpus link for the original license; commercial use may require verifying the third-party API terms for agent, switch, chat, lmChatGoogleGemini, memoryBufferWindow, embeddingsGoogleGemini, summarize, vectorStoreQdrant.