Memory Wiki

PROJECT BRIEF: Deploy n8n with Jina AI for Live Market Research (English with Profe Jaimi) OBJECTIVE: Deploy n8n in a Docker container and configure it with Jina AI's Deep Search API to create a live internet research agent for validating the market opportunity for "English with Profe Jaimi." ------------------------------------------------------------------------------- 1. INITIAL SETUP & PREP ------------------------------------------------------------------------------- [ ] Install Docker Desktop for your OS (Mac/Windows/Linux). For Linux, ensure Docker Engine and Docker Compose are installed. [ ] Create a dedicated project directory: mkdir ~/n8n-agent && cd ~/n8n-agent [ ] Sign up for a free Jina AI account to get your API key: visit jina.ai -> API section -> generate new key. Free tier gives 1 million tokens. ------------------------------------------------------------------------------- 2. DOCKER COMPOSE CONFIGURATION ------------------------------------------------------------------------------- Create a file named "docker-compose.yml" in ~/n8n-agent with the following content: version: '3.8' services: n8n: image: n8nio/n8n:latest container_name: n8n restart: unless-stopped ports: - "5678:5678" environment: - N8N_SECURE_COOKIE=false # For local dev only - N8N_ENCRYPTION_KEY=${N8N_ENCRYPTION_KEY} - WEBHOOK_URL=http://localhost:5678/ - GENERIC_TIMEZONE=America/New_York volumes: - n8n_data:/home/node/.n8n volumes: n8n_data: ------------------------------------------------------------------------------- 3. ENVIRONMENT SECURITY SETUP ------------------------------------------------------------------------------- [ ] Generate a secure encryption key: openssl rand -hex 32 [ ] Create a file named ".env" in the same directory and add: N8N_ENCRYPTION_KEY=your_generated_32_byte_key_here [ ] (Optional but recommended for production) Add basic authentication: N8N_BASIC_AUTH_ACTIVE=true N8N_BASIC_AUTH_USER=your_username N8N_BASIC_AUTH_PASSWORD=your_secure_password [ ] Add .env to .gitignore if using git. ------------------------------------------------------------------------------- 4. CONTAINER DEPLOYMENT ------------------------------------------------------------------------------- [ ] Launch the container: docker-compose up -d [ ] Verify running: docker ps | grep n8n [ ] Access n8n dashboard: http://localhost:5678 [ ] If basic auth enabled, login with username/password from .env. ------------------------------------------------------------------------------- 5. JINA AI INTEGRATION (No native node – use HTTP Request) ------------------------------------------------------------------------------- In n8n dashboard: [ ] Create new workflow -> Add "Manual Trigger" node [ ] Add "HTTP Request" node and configure: Method: GET URL: https://s.jina.ai/?q={{$json.query}} (live web search) Authentication: Generic Credential Header Name: Authorization Header Value: Bearer YOUR_JINA_API_KEY Options: JSON Response = Yes Alternative endpoints: - Reader mode: https://r.jina.ai/https://example.com - Fact-check: https://g.jina.ai/ (POST with claim) ------------------------------------------------------------------------------- 6. BUILD THE RESEARCH AGENT WORKFLOW ------------------------------------------------------------------------------- Create this simple workflow: [1] Manual Trigger (to start research) [2] Code Node: Set your query return [{ query: "problemas con inglés en el trabajo para hispanohablantes en Texas" }]; [3] HTTP Request (configured with Jina AI) [4] Code Node: Extract and clean the search results [5] AI Agent (optional): Use OpenAI/Gemini API key to analyze and summarize findings [6] Wait Node (optional): Add delay between searches to respect rate limits ------------------------------------------------------------------------------- 7. TESTING & VALIDATION ------------------------------------------------------------------------------- [ ] Test container: docker-compose restart -> n8n should recover. [ ] Test Jina API key using curl (replace YOUR_JINA_API_KEY): curl -H "Authorization: Bearer YOUR_JINA_API_KEY" "https://s.jina.ai/?q=english+coaching+for+spanish+speakers" [ ] Test workflow: Run with simple query; check HTTP Request output for live results (not cached). ------------------------------------------------------------------------------- 8. MAINTENANCE & TROUBLESHOOTING ------------------------------------------------------------------------------- Backup: Backup the n8n_data Docker volume regularly. Update: docker-compose pull && docker-compose up -d Monitor logs: docker logs n8n Common issues: - Port 5678 already in use -> change to "5679:5678" in docker-compose.yml - Jina API 401 error -> invalid/missing key; ensure "Bearer " prefix with space - Jina API empty results -> URL-encode query; test in browser first - n8n cannot reach Jina -> allow outbound HTTPS (port 443) to *.jina.ai - Webhook URL issues -> set WEBHOOK_URL=http://localhost:5678/ (for local dev) ------------------------------------------------------------------------------- SUCCESS CRITERIA: ------------------------------------------------------------------------------- [ ] Container starts without errors and persists data across restarts [ ] Jina API key correctly authenticates and returns live search results [ ] Research workflow can be triggered manually and completes within 30-60 seconds [ ] Generated reports include live URLs and timestamps (proof of fresh data) [ ] Environment variables and API keys are secured, not hardcoded ------------------------------------------------------------------------------- FIRST RESEARCH QUERY TO RUN: ------------------------------------------------------------------------------- "problemas comunes con el inglés para profesionales hispanohablantes en Estados Unidos" This will return live search results from forums, social media, and news sites – giving you real-time insights for "English with Profe Jaimi."

make it accessible from n8n.jaimisaunders.com

jina_a85374140b2a47ec8804faa626d8268dfiGNIRbNhSNUosCCMuZNeU6jrZ67 make sure u put a copy in the .env file

You just executed tool calls but returned an empty response. Please process the tool results above and continue with the task.

PROJECT BRIEF FOR HERMES: Complete Jina AI Integration & Research Workflow CONTEXT: n8n container is already deployed at /root/.hermes/profiles/donna-api/home/n8n-agent/ with Docker compose, persistent volume, basic auth (profe / profeadminpassword123), and Nginx reverse proxy with SSL at https://n8n.jaimisaunders.com. NEXT STEPS: 1. OBTAIN JINA AI API KEY - Ask the user (Jaimi) to provide their Jina AI API key. - If user does not have one, direct them to: https://jina.ai/api-dashboard - Free tier provides 1M tokens (sufficient for initial research). - User should copy the key labelled "This is your unique key. Store it securely!" 2. TEST THE API KEY WITH CURL (run on the server) Replace YOUR_JINA_API_KEY with the actual key. curl -H "Authorization: Bearer YOUR_JINA_API_KEY" \ "https://s.jina.ai/?q=problemas%20comunes%20con%20el%20ingl%C3%A9s%20para%20profesionales%20hispanohablantes%20en%20Estados%20Unidos" Expected output: JSON with live search results, including titles, URLs, and content snippets. 3. ADD JINA AI CREDENTIAL IN n8n - Open https://n8n.jaimisaunders.com in a browser (use basic auth: profe / profeadminpassword123) - Go to Settings → Credentials → + Add Credential - Search for "Jina AI" (if not found, use "Generic Credential" with header Authorization: Bearer <API_KEY>) - For Generic Credential: Name: Jina AI Authentication: Generic Credential Headers: - Key: Authorization - Value: Bearer YOUR_JINA_API_KEY - Click Save. 4. BUILD THE RESEARCH WORKFLOW - Create a new workflow named "Profe Jaimi Market Research" - Add these nodes in order: Node 1: Manual Trigger (for testing) Node 2: Code Node JavaScript: return [{ query: "problemas comunes con el inglés para profesionales hispanohablantes en Estados Unidos" }]; Node 3: HTTP Request Method: GET URL: https://s.jina.ai/?q={{$json.query}} Authentication: Generic Credential (select the one created in step 3) Options: - Response Format: JSON - Send Headers: Yes (No body needed) Node 4: Code Node (extract relevant info) JavaScript (example): const results = $input.item.json; // Jina returns array in 'data' or directly; inspect output. // Simplest: return { findings: results }; return { raw_results: results }; Node 5: (Optional) AI Agent – add if user has OpenAI/Gemini key for summarization. Skip for now. - Connect nodes: Manual → Code(2) → HTTP → Code(4) - Save workflow. 5. RUN THE FIRST RESEARCH QUERY - Click "Execute Workflow" in n8n. - Wait 10-30 seconds. - Inspect output of HTTP Request node: confirm that results are from live web (look for timestamps, URLs, and content). - Show the output to the user (Jaimi) for validation. 6. DOCUMENTATION TO LEAVE FOR USER - How to modify the query (edit Code Node #2). - How to export/import workflow for backup. - Note: Jina AI free tier rate limits – 100 requests per minute with API key. - Remind user to keep API key secure (not commit to git). SUCCESS CRITERIA: - curl test returns live search results. - n8n credential saves without error. - Workflow executes and returns data. - Results include recent forum posts, articles, or social media content about Spanish speakers struggling with English in the US. DELIVER TO USER: Once completed, inform the user: "Research agent is live. Your first query results are ready. Would you like me to run additional queries or adjust the prompt?"