Gathering detailed insights and metrics for @agenite/bedrock
🤖 Build powerful AI agents with TypeScript. Agenite makes it easy to create, compose, and control AI agents with first-class support for tools, streaming, and multi-agent architectures. Switch seamlessly between providers like OpenAI, Anthropic, AWS Bedrock, and Ollama.
Installations
npm install @agenite/bedrockDeveloper Guide
BETA
Typescript
Yes
Module System
ESM
Node Version
22.13.0
NPM Version
11.1.0
Contributors
2
Unable to fetch Contributors
Languages
4
TypeScript
CSS
JavaScript
Handlebars
TypeScript (92.04%)
CSS (5.27%)
JavaScript (2.56%)
Handlebars (0.13%)
validate.email 🚀
Verify real, reachable, and deliverable emails with instant MX records, SMTP checks, and disposable email detection.
Developer
subeshb1
Download Statistics
Total Downloads
939
Last Day
30
Last Week
42
Last Month
371
Last Year
939
GitHub Statistics
MIT License
30 Stars
70 Commits
2 Forks
1 Watchers
2 Branches
2 Contributors
Updated on Mar 12, 2025
Bundle Size
159.56 kB
Minified
46.99 kB
Minified + Gzipped
Maintainers
1
Package Meta Information
Latest Version
0.3.1
Package Id
@agenite/bedrock@0.3.1
Unpacked Size
98.77 kB
Size
15.00 kB
File Count
10
NPM Version
11.1.0
Node Version
22.13.0
Published on
Feb 25, 2025
Total Downloads
Cumulative downloads
Total Downloads
939
Last Day
500%
30
Compared to previous day
Last Week
-30%
42
Compared to previous week
Last Month
-34.7%
371
Compared to previous month
Last Year
0%
939
Compared to previous year
Daily Downloads
Weekly Downloads
Monthly Downloads
Yearly Downloads
@agenite/bedrock
AWS Bedrock provider for Agenite, enabling seamless integration with Amazon's foundation models through the Bedrock runtime.
Features
- 🚀 Easy Integration - Simple setup with AWS Bedrock
- 🔄 Streaming Support - Real-time streaming responses
- 🛠️ Tool Integration - Native support for function calling
- 🎯 Model Selection - Support for multiple Bedrock models
- 🔐 AWS Authentication - Automatic credential management
- 🌐 Region Support - Configurable AWS regions
Installation
1npm install @agenite/bedrock
Quick Start
1import { BedrockProvider } from '@agenite/bedrock'; 2 3// Initialize the provider 4const provider = new BedrockProvider({ 5 model: 'anthropic.claude-3-5-haiku-20241022-v1:0', 6 region: 'us-west-2', 7}); 8 9// Generate a simple response 10const result = await provider.generate( 11 'What are the main features of Llama 2?' 12); 13console.log(result); 14 15// Use streaming for real-time responses 16const generator = provider.stream('Tell me about AWS Bedrock.'); 17for await (const chunk of generator) { 18 if (chunk.type === 'text') { 19 process.stdout.write(chunk.text); 20 } 21}
Configuration
Provider Options
1interface BedrockProviderOptions { 2 model: string; // Bedrock model ID 3 region: string; // AWS region 4 temperature?: number; // Temperature for response generation 5 maxTokens?: number; // Maximum tokens in response 6 topP?: number; // Top P sampling parameter 7}
Available Models
- Anthropic Claude Models:
anthropic.claude-3-5-haiku-20241022-v1:0anthropic.claude-3-sonnet-20240229-v1:0anthropic.claude-instant-v1
- Amazon Titan Models:
amazon.titan-text-express-v1amazon.titan-text-lite-v1
Advanced Usage
Usage with Claude 3.7 thinking
1const provider = new BedrockProvider({ 2 model: `us.anthropic.claude-3-7-sonnet-20250219-v1:0`, 3 region: 'us-east-2', 4 converseCommandConfig: { 5 additionalModelRequestFields: { 6 reasoning_config: { 7 type: 'enabled', 8 budget_tokens: 1024, 9 }, 10 }, 11 inferenceConfig: { 12 temperature: 1, 13 }, 14 }, 15});
Tool Integration
1import { BedrockProvider } from '@agenite/bedrock'; 2import type { ToolDefinition } from '@agenite/llm'; 3 4// Define a calculator tool 5const calculatorTool: ToolDefinition = { 6 name: 'calculator', 7 description: 'Perform basic arithmetic operations', 8 inputSchema: { 9 type: 'object', 10 properties: { 11 operation: { 12 type: 'string', 13 enum: ['add', 'subtract', 'multiply', 'divide'], 14 }, 15 a: { type: 'number' }, 16 b: { type: 'number' }, 17 }, 18 required: ['operation', 'a', 'b'], 19 }, 20}; 21 22// Initialize provider with tool support 23const provider = new BedrockProvider({ 24 model: 'anthropic.claude-3-5-haiku-20241022-v1:0', 25 region: 'us-west-2', 26}); 27 28// Use tool in conversation 29const messages = [ 30 { 31 role: 'user', 32 content: [{ type: 'text', text: 'What is 123 multiplied by 456?' }], 33 }, 34]; 35 36const generator = provider.iterate(messages, { 37 tools: [calculatorTool], 38 stream: true, 39 systemPrompt: 40 'You are a helpful AI assistant with access to a calculator tool.', 41}); 42 43// Process streaming response with tool usage 44for await (const chunk of generator) { 45 if (chunk.type === 'text') { 46 process.stdout.write(chunk.text); 47 } else if (chunk.type === 'toolUse') { 48 console.log('Tool Use:', chunk); 49 } 50}
Conversation Management
1// Maintain conversation history 2let messages = []; 3const result = await provider.generate(messages, { 4 systemPrompt: 'You are a helpful AI assistant.', 5}); 6messages.push( 7 { role: 'user', content: [{ type: 'text', text: 'Hello!' }] }, 8 result 9);
API Reference
BedrockProvider Class
1class BedrockProvider implements LLMProvider { 2 constructor(options: BedrockProviderOptions); 3 4 generate( 5 messages: string | BaseMessage[], 6 options?: GenerateOptions 7 ): Promise<BaseMessage>; 8 9 stream( 10 messages: string | BaseMessage[], 11 options?: StreamOptions 12 ): AsyncGenerator<StreamChunk>; 13 14 iterate( 15 messages: string | BaseMessage[], 16 options?: StreamOptions 17 ): AsyncGenerator<StreamChunk>; 18}
Message Types
1interface BaseMessage { 2 role: 'user' | 'assistant' | 'system'; 3 content: ContentBlock[]; 4} 5 6type ContentBlock = TextBlock | ToolUseBlock | ToolResultBlock;
Examples
Check out the examples directory for more:
basic-chat.ts- Simple chat interactiontool-usage.ts- Advanced tool integration example
AWS Setup
- Install AWS CLI and configure credentials:
1aws configure
-
Ensure your AWS account has access to Bedrock and the required models.
-
Set up IAM permissions for Bedrock access:
1{ 2 "Version": "2012-10-17", 3 "Statement": [ 4 { 5 "Effect": "Allow", 6 "Action": [ 7 "bedrock:InvokeModel", 8 "bedrock:InvokeModelWithResponseStream" 9 ], 10 "Resource": "*" 11 } 12 ] 13}
Contributing
Contributions are welcome! Please feel free to submit a Pull Request.
License
MIT
No vulnerabilities found.
No security vulnerabilities found.