Gathering detailed insights and metrics for @aurogpt/flowgpt
Gathering detailed insights and metrics for @aurogpt/flowgpt
FlowGPT is a framework for creating AI bots that follow predefined step sequences.
npm install @aurogpt/flowgpt
Typescript
Module System
Node Version
NPM Version
62.3
Supply Chain
91.1
Quality
83.9
Maintenance
100
Vulnerability
98.2
License
TypeScript (96.67%)
JavaScript (3.33%)
Love this project? Help keep it running — sponsor us today! 🚀
Total Downloads
605
Last Day
2
Last Week
12
Last Month
71
Last Year
605
MIT License
30 Commits
1 Forks
1 Watchers
2 Branches
2 Contributors
Updated on Jan 05, 2025
Minified
Minified + Gzipped
Latest Version
2.1.3
Package Id
@aurogpt/flowgpt@2.1.3
Unpacked Size
938.79 kB
Size
181.72 kB
File Count
28
NPM Version
10.8.2
Node Version
18.20.5
Published on
Jan 05, 2025
Cumulative downloads
Total Downloads
Last Day
100%
2
Compared to previous day
Last Week
-40%
12
Compared to previous week
Last Month
-86.7%
71
Compared to previous month
Last Year
0%
605
Compared to previous year
7
1
6
FlowGPT is an enterprise-grade framework for developing conversational AI bots with structured dialogue flows. Built on LangChain, it provides seamless integration with OpenAI and Azure OpenAI models, enabling developers to create sophisticated conversation scenarios with predefined step sequences.
1npm install @aurogpt/flowgpt
The following example demonstrates how to create a basic customer service bot with product browsing and order tracking capabilities:
1import { FlowBot, Scenario } from "@aurogpt/flowgpt"; 2import { ChatOpenAI } from "langchain/chat_models/openai"; 3 4const flowBot = FlowBot.create({ 5 name: "CustomerServiceBot", 6 description: "Automated customer service assistant", 7 scenarios: [ 8 Scenario.loadObject({ 9 name: "Browse Products", 10 description: "Helps customers find and browse products", 11 steps: [ 12 { 13 name: "Greet User", 14 description: "Welcome the user to the store", 15 instructions: 16 "Send a greeting message with options like 'View Products' or 'Search for a Product'.", 17 }, 18 { 19 name: "Product Search", 20 description: "Assist the user in finding specific products", 21 instructions: 22 "Ask for the product name or category and display matching results.", 23 }, 24 { 25 name: "View Product Details", 26 description: 27 "Provide detailed information about a selected product", 28 instructions: 29 "Show product details, price, and an 'Add to Cart' option.", 30 }, 31 ], 32 }), 33 Scenario.loadObject({ 34 name: "Track Order", 35 description: "Allows customers to check the status of their order", 36 steps: [ 37 { 38 name: "Request Order ID", 39 description: "Ask the user to provide their order ID", 40 instructions: 41 "Prompt the user for their order ID to retrieve tracking information.", 42 }, 43 { 44 name: "Display Order Status", 45 description: "Show the current status of the order", 46 instructions: 47 "Provide shipping updates, expected delivery date, and tracking details.", 48 }, 49 ], 50 }), 51 ], 52 model: new ChatOpenAI({ 53 modelName: "gpt-4o-mini", 54 // apiKey: process.env.OPENAI_API_KEY, 55 }), 56 57 // Optional: Prompt for detecting scenario & step 58 // prompt: "...", 59}); 60 61// Example extending default prompt 62flowBot.prompt += "This text is appended to the default prompt"; 63 64// Example usage 65const response = await flowBot.invoke(` 66 User: Hi 67 Bot: Hello, How can I help you today? 68 User: What is the status of my order id #324? 69`); 70 71console.log(response);
The FlowBot
constructor accepts the following configuration options:
name
: Bot identifierdescription
: Bot purpose and functionality descriptionscenarios
: Array of conversation scenariosmodel
: LLM instance for detecting scenario & stepFor detailed examples and advanced configuration options, please refer to our documentation.
We welcome contributions from the community. To contribute:
Please review our Contributing Guidelines for more information.
FlowGPT is released under the MIT License.
No vulnerabilities found.
No security vulnerabilities found.