yusan25c
Workflows by yusan25c
Create knowledge base from Jira tickets with OpenAI embeddings and Pinecone
### **How It Works** This template is a workflow that **registers Jira tickets to Pinecone**. By combining it with the [Automated Jira Ticket Responses with GPT-4 and Pinecone Knowledge Base](https://creators.n8n.io/workflows/9087) template, you can continuously improve the quality of automated responses in Jira. --- ### **Prerequisites** - A Jira account and credentials (API key and email address) - A Pinecone account and credentials (API key and environment settings) - OpenAI credentials (API key) --- ### **Setup Instructions** 1. **Jira Credentials** - Register your Jira credentials (API key and email address) in n8n. 2. **Vector Database Setup (Pinecone)** - Register your Pinecone credentials (API key and environment variables) in n8n. 3. **AI Node** - Configure the OpenAI node with your credentials (API key). --- ### **Step by Step** 1. **Scheduled Trigger** The workflow runs at regular intervals according to the schedule set in the Scheduled Trigger node. 2. **Jira Trigger (Completed Tickets)** Retrieves the summary, description, and comments of completed Jira tickets. 3. **Register to Pinecone** Converts the retrieved ticket information into vectors and registers them in Pinecone. --- ### **Notes** - Configure the Scheduled Trigger interval carefully to avoid exceeding API rate limits. --- ### **Further Reference** For a detailed walkthrough (in Japanese), see this article: 👉 [Automating knowledge registration to Pinecone with n8n (Qiita)](https://qiita.com/yusan25c/items/f8d0cca479310af171a6) You can find the template file on GitHub here: 👉 [Template File on GitHub](https://github.com/yusan25c/n8n-jira-knowledge-sync-template/tree/main/02_Jira_Vector-Register-Minimal)
Automated Jira ticket responses with GPT-4 and Pinecone knowledge base
### **How It Works** This template is an **n8n workflow that integrates with Jira to provide automated replies**. When a ticket is assigned to a user, the workflow analyzes the ticket content, retrieves relevant knowledge from a vector database, and generates a response. By continuously enriching the knowledge base, the system improves response quality in Jira. --- ### **Prerequisites** - A Jira account with API access - A Pinecone account and credentials (API key and environment settings) - An AI provider credential (e.g., OpenAI API key) --- ### **Setup Instructions** 1. **Jira Credentials** - Create Jira credentials in n8n (API token and email). - In the Jira node, **select the registered Jira account ID**. 2. **Vector Database Setup (Pinecone)** - Register your **Pinecone credentials (API key and environment variables)** in n8n. - Ensure that your knowledge base is indexed in Pinecone. 3. **AI Assistant Node** - Configure the OpenAI (or other LLM) node with your API key. - Provide a system prompt that explains how to respond to Jira tickets using retrieved knowledge. 4. **Workflow Execution** - The workflow runs **only via the Scheduled Trigger node** at defined intervals. - When Jira tickets are assigned, their summary, description, and latest comments are retrieved. - These details are passed to the AI assistant, which queries Pinecone and generates a response. - The generated response is then posted as a Jira comment. --- ### **Step by Step** 1. **Scheduled Trigger** The workflow is executed at regular intervals using the Scheduled Trigger node. 2. **Jira Trigger (Issue Assigned)** Retrieves the summary, description, and latest comments of assigned tickets. 3. **AI Assistant** Sends ticket details to the AI assistant, which searches and summarizes relevant knowledge from Pinecone. 4. **Response Generation / Ticket Update** The AI generates a response and automatically posts it as a Jira comment. (Optionally, the workflow can update the ticket status or mention the assignee.) --- ### **Notes** - Keep your Pinecone knowledge base updated to improve accuracy. - You can customize the AI assistant’s behavior by adjusting the system prompt. - Configure the Scheduled Trigger frequency carefully to avoid API rate limits. --- ### **Further Reference** For a detailed walkthrough (in Japanese), see this article: 👉 [Automating Jira responses with n8n, AI, and Pinecone (Qiita)](https://qiita.com/yusan25c/items/f6b15aef451556f4c9fe) You can find the template file on GitHub here: 👉 [Template File on GitHub](https://github.com/yusan25c/n8n-jira-knowledge-sync-template/tree/main/01_Jira_Knowledge-Sync-Minimal-Summary-Comment-Assignee)