> For the complete documentation index, see [llms.txt](https://docs.snackprompt.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.snackprompt.com/bring-your-data-into-ai/how-to/get-started-with-integrations/how-to-integrate-with-n8n.md). # How to Integrate with N8N Learn how to use the SnackPrompt AI Engine API as a knowledge source for AI agents and workflows in N8N. ### Overview N8N offers several ways to integrate external APIs with its AI capabilities: | Method | Use Case | Description | | --------------------- | ----------------- | ----------------------------------------------- | | **HTTP Request Tool** | Agents with tools | Agent decides when to query the API | | **HTTP Request Node** | RAG in workflows | Direct call at a specific point in the workflow | | **Custom Retriever** | Advanced RAG | Replace native vector store with external API | ### Integration Architecture ``` ┌─────────────────────────────────────────────────────────┐ │ N8N │ │ ┌─────────────┐ ┌─────────────┐ ┌────────────┐ │ │ │ Trigger │───▶│ AI Agent │───▶│ Response │ │ │ └─────────────┘ └──────┬──────┘ └────────────┘ │ │ │ │ │ ┌──────▼──────┐ │ │ │ HTTP Tool │ │ │ └──────┬──────┘ │ └────────────────────────────┼────────────────────────────┘ │ ▼ ┌──────────────────────────────┐ │ SnackPrompt AI Engine API │ │ /v1/kb/search or /v1/kb/chat│ └──────────────────────────────┘ ``` *** ### Method 1: HTTP Request Tool (Recommended) Use when you want the **agent to autonomously decide** when to search your knowledge base. #### Step 1: Configure the AI Agent 1. Add an **AI Agent** node to your workflow 2. Configure the LLM model (OpenAI, Anthropic, etc.) 3. Connect an **HTTP Request Tool** as a tool #### Step 2: Configure the HTTP Request Tool **Tool Settings:** | Field | Value | | ----------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | Name | `search_knowledge_base` | | Description | `Use this tool to search for information in the company knowledge base. Send a natural language query to find relevant documents about products, policies, procedures, etc.` | | Method | `POST` | | URL | `https://api-integrations.snackprompt.com/v1/kb/search` | **Headers:** ``` Content-Type: application/json x-api-key: YOUR_API_KEY ``` **Body (JSON):** ```json { "query": "{{ $fromAI('query', 'The search query') }}", "filters": { "tenant_id": "YOUR_TENANT_ID" }, "limit": 5 } ``` #### Step 3: Optimize the Response In the HTTP Request Tool **Options** section: * **Optimize Response**: Enabled * **Response Format**: `JSON` * **Limit Response Size**: Recommended to reduce tokens #### Complete Workflow Example ``` [Chat Trigger] → [AI Agent] → [Response] │ └─── [HTTP Request Tool: search_knowledge_base] ``` The agent will receive a question, decide if it needs to query the knowledge base, perform the search, and use the results to formulate the response. *** ### Method 2: HTTP Request Node (Direct RAG) Use when you want a **deterministic workflow** where the search always happens. #### Simple RAG Workflow ``` [Webhook/Chat Trigger] │ ▼ [HTTP Request: Search API] │ ▼ [Set Node: Format Context] │ ▼ [AI Chain: Generate Response] │ ▼ [Respond to Webhook] ``` #### HTTP Request Node Configuration **Method:** POST **URL:** ``` https://api-integrations.snackprompt.com/v1/kb/search ``` **Headers:** ``` Content-Type: application/json x-api-key: YOUR_API_KEY ``` **Body:** ```json { "query": "{{ $json.chatInput }}", "filters": { "tenant_id": "YOUR_TENANT_ID" }, "limit": 5 } ``` #### Formatting the Context Use a **Set Node** or **Code Node** to format the results: ```javascript // Code Node const results = $input.first().json.items; const context = results.map((item, index) => `[${index + 1}] ${item.payload.original_text}` ).join('\n\n'); return { context: context, sources: results.map(r => ({ id: r.payload.snack_item_id, score: r.score })) }; ``` #### AI Chain Prompt ``` You are a helpful assistant. Answer the user's question based ONLY on the following context. Context: {{ $json.context }} User Question: {{ $('Webhook').item.json.chatInput }} If the context doesn't contain relevant information, say "I don't have information about that." ``` *** ### Method 3: Chat Endpoint for Complete Responses Use the `/v1/kb/chat` endpoint when you want the **API to handle all the RAG** and return a ready response. #### Configuration **URL:** ``` https://api-integrations.snackprompt.com/v1/kb/chat ``` **Headers:** ``` Content-Type: application/json x-api-key: YOUR_API_KEY ``` **Body:** ```json { "query": "{{ $json.chatInput }}", "filters": { "tenant_id": "YOUR_TENANT_ID", "tag_names": ["Support", "FAQ"] } } ``` #### Response The API returns: * `answer`: AI-generated response * `sources`: Sources used to generate the response You can use it directly or enrich with additional logic in N8N. *** ### Practical Use Cases #### 1. Support Chatbot ``` [Chat Trigger] → [HTTP Request: /chat] → [Respond with answer] ``` Ideal for simple chatbots that need to answer questions about products, policies, etc. #### 2. Multi-tool Agent ``` [Chat Trigger] → [AI Agent] → [Response] │ ├── [Tool: search_knowledge_base] ├── [Tool: Google Calendar] └── [Tool: Send Email] ``` The agent can search for information AND execute actions. #### 3. RAG with Multiple Sources ``` [Webhook] → [Switch: Route by Topic] │ ├── [Search: tag=Sales] → [Merge] → [AI: Generate] └── [Search: tag=Support] ───┘ ``` Search in different categories and combine the results. #### 4. Response Validation ``` [Chat] → [AI Agent: Generate Draft] │ ▼ [HTTP Request: Search for validation] │ ▼ [AI: Verify and Refine] │ ▼ [Response] ``` The agent generates a response, searches for validation in the knowledge base, and refines. *** ### Configuration Tips #### 1. Tool Description is Crucial The HTTP Request Tool description determines **when** the agent will use it: ``` ✅ Good description: "Search the company knowledge base for information about products, pricing, policies, and procedures. Use when the user asks about company-specific information." ❌ Bad description: "Search API" ``` #### 2. Limit the Results Too many results = too many tokens = higher cost and possible confusion: ```json { "limit": 3 // Start with few and increase if needed } ``` #### 3. Use Filters for Context Direct the search with tags when you know the context: ```json { "filters": { "tenant_id": "...", "tag_names": ["{{ $json.detected_topic }}"] } } ``` #### 4. Handle Errors Add an **Error Trigger** or use **Continue On Fail** to handle API failures. *** ### Complete Example: Sales Chatbot #### Workflow 1. **Chat Trigger**: Receives user message 2. **AI Agent**: Processes with GPT-4 3. **HTTP Request Tool**: Searches products and prices 4. **HTTP Request Tool**: Searches discount policies 5. **Response**: Returns response to user #### Agent Configuration **System Prompt:** ``` You are a sales assistant for company X. Help customers find products and understand pricing. Use the search_products tool to find product information. Use the search_policies tool to find discount policies. Always cite sources when providing information. ``` **Tools:** | Tool | Description | | ----------------- | ----------------------------------------------------------- | | `search_products` | Search for product information, specifications, and pricing | | `search_policies` | Search for discount policies, payment terms, and conditions | *** ### Troubleshooting #### Error: "tenant\_id is required" Make sure `tenant_id` is inside the `filters` object: ```json // ❌ Wrong { "query": "...", "tenant_id": "..." } // ✅ Correct { "query": "...", "filters": { "tenant_id": "..." } } ``` #### Agent doesn't use the tool 1. Improve the tool description 2. Add examples in the system prompt 3. Verify if the question actually requires the tool #### Response too long/truncated 1. Reduce the results `limit` 2. Enable **Optimize Response** on the tool 3. Use a Code Node to summarize before passing to the LLM *** ### Related * [Endpoints Reference](/bring-your-data-into-ai/reference/endpoints.md) * [Available Filters](/bring-your-data-into-ai/reference/filters.md) * [Error Handling](/bring-your-data-into-ai/how-to/how-to-handle-errors.md) ### External Resources * [N8N RAG Documentation](https://docs.n8n.io/advanced-ai/rag-in-n8n/) * [HTTP Request Tool Docs](https://docs.n8n.io/integrations/builtin/cluster-nodes/sub-nodes/n8n-nodes-langchain.toolhttprequest/) * [Building RAG in 2025 - N8N Community](https://community.n8n.io/t/building-rag-in-2025-vector-stores-as-tools-is-here/75166) --- # Agent Instructions This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com. ## Querying This Documentation If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question. Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter: ``` GET https://docs.snackprompt.com/bring-your-data-into-ai/how-to/get-started-with-integrations/how-to-integrate-with-n8n.md?ask=&goal= ``` `ask` is the immediate question: it should be specific, self-contained, and written in natural language. `goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal. The response will contain a direct answer to the question and relevant excerpts and sources from the documentation. Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.