Merge pull request #134 from mongodb-developer/openai-agents-sdk-js

OpenAI Agents SDK for TypeScript

Author: codeSTACKr

apps/README.md

@@ -6,6 +6,7 @@ Javascript and Python apps and demos showcasing how to use MongoDB in GenAI appl
 | Local RAG PDF | MongoDB Atlas, Python | Local RAG implementation for PDF processing and querying | [![View Code](https://img.shields.io/badge/view-code-blue?logo=github)](local-rag-pdf) |
 | Lyric Semantic Search | MongoDB Atlas, Spring AI, Java | Semantic search service for song lyrics with metadata filtering using Spring AI framework | [![View Code](https://img.shields.io/badge/view-code-blue?logo=github)](lyric-semantic-search) |
 | MongoMP | MongoDB Atlas, Next.js 14, Node.js | Music streaming platform with AI-powered recommendations using vector search and real-time tracking | [![View Code](https://img.shields.io/badge/view-code-blue?logo=github)](mongo-mp) |
+| OpenAI Agents SDK | MongoDB Atlas, TypeScript, OpenAI Agents SDK | Template for using OpenAI Agents SDK with MongoDB as a memory store for conversation history | [![View Code](https://img.shields.io/badge/view-code-blue?logo=github)](openai-agents-sdk) |
 | RT Voice TS Store Agent | MongoDB Atlas, Next.js, WebRTC, OpenAI | Voice-controlled e-commerce platform with real-time voice processing and hybrid search | [![View Code](https://img.shields.io/badge/view-code-blue?logo=github)](RT-voice-ts-store-agent) |
 | SpringAI Terraform RAG | MongoDB Atlas, Spring Boot, Terraform, OpenAI | RAG application with automated infrastructure setup using Terraform and Spring Boot | [![View Code](https://img.shields.io/badge/view-code-blue?logo=github)](springai-terraform-rag) |
 | Voice OpenAI Rentals Agent | MongoDB Atlas, Python, Flask, OpenAI | Voice interaction agent for real estate listings using vector search and real-time voice processing | [![View Code](https://img.shields.io/badge/view-code-blue?logo=github)](voice-openai-mongo-rentals-agent) |

apps/openai-agents-sdk/.env.example

@@ -0,0 +1,2 @@
+MONGODB_URI=your_connection_string
+OPENAI_API_KEY=sk-...

apps/openai-agents-sdk/.gitignore

@@ -0,0 +1,55 @@
+# Dependency directories
+node_modules/
+jspm_packages/
+
+# Build outputs
+dist/
+build/
+out/
+.next/
+.nuxt/
+
+# Coverage directories
+coverage/
+.nyc_output/
+
+# Environment variables
+.env
+.env.local
+.env.development.local
+.env.test.local
+.env.production.local
+
+# Logs
+logs/
+*.log
+npm-debug.log*
+yarn-debug.log*
+yarn-error.log*
+lerna-debug.log*
+
+# Cache directories
+.npm/
+.eslintcache
+.stylelintcache
+.rpt2_cache/
+.rts2_cache_cjs/
+.rts2_cache_es/
+.rts2_cache_umd/
+
+# Editor directories and files
+.idea/
+.vscode/
+*.swp
+*.swo
+.DS_Store
+Thumbs.db
+
+# TypeScript cache
+*.tsbuildinfo
+
+# Next.js
+.vercel
+
+# MongoDB
+*.pem

apps/openai-agents-sdk/README.md

@@ -0,0 +1,71 @@
+# OpenAI Agents SDK with MongoDB Memory
+
+This demo app shows how to use the [OpenAI Agents SDK](https://openai.github.io/openai-agents-js/) with [MongoDB](https://www.mongodb.com/cloud/atlas/register/?utm_campaign=devrel&utm_source=third-party-content&utm_medium=cta&utm_content=openai-agents-sdk-template&utm_term=jesse.hall) for persistent conversation history.
+
+## Key Features
+
+- Uses MongoDB to store conversation history
+- Leverages the SDK's built-in message handling capabilities
+- Custom tool to access user profile information
+- Simple implementation for educational purposes
+
+## Getting Started
+
+1. Download this app folder using:
+   ```bash
+   npx degit mongodb-developer/GenAI-Showcase/apps/openai-agents-sdk my-openai-agent
+   ```
+
+2. Make sure MongoDB is set up and running
+
+3. Copy the example environment file and add your credentials:
+   ```bash
+   cp .env.example .env
+   ```
+   Then update the `.env` file with your information:
+   ```
+   MONGODB_URI=mongodb://localhost:27017/agent_memory
+   OPENAI_API_KEY=your_openai_api_key
+   ```
+
+4. Install dependencies:
+   ```bash
+   npm install
+   ```
+
+5. Build the TypeScript:
+   ```bash
+   npm run build
+   ```
+
+6. Run the demo:
+   ```bash
+   npm start
+   ```
+
+## How It Works
+
+The demo shows a hybrid approach to conversation memory:
+
+1. **MongoDB for Persistence**: Messages are saved to MongoDB so they persist across server restarts
+2. **SDK's Message Handling**: We use the SDK's built-in `user()` and `assistant()` functions to format message history
+3. **Context-Aware Tools**: Custom tools can access application context including user profiles
+4. **Simple Interface**: Clean implementation with minimal code
+
+## Demo Conversation
+
+The demo runs a simple conversation that demonstrates:
+1. Basic question answering
+2. Memory of previous questions (asking about "its population" after asking about France)
+3. Using a tool to fetch user profile information
+
+## Project Structure
+
+- `src/db.ts`: MongoDB connection and conversation storage functions
+- `src/agent.ts`: OpenAI Agents SDK implementation with custom tools
+- `src/types.ts`: TypeScript type definitions for conversations and context
+- `src/index.ts`: Demo script that runs a sample conversation
+
+## License
+
+MIT