bedrock-wrapper 2.4.4 → 2.5.0

package/CHANGELOG.md

@@ -1,6 +1,37 @@
 # Changelog
 All notable changes to this project will be documented in this file.
 
+## [2.5.0] - 2025-08-12 (Converse API)
+### Added
+- Support for Converse API (streaming and non-streaming)
+
+### Technical Details
+- **Model Configuration**: All models use standard messages API format
+- **API Compatibility**: Supports OpenAI-style requests
+- **Response Processing**: Automatic reasoning tag handling based on model variant
+- **Streaming Fallback**: Automatic detection and fallback to non-streaming for unsupported models
+- **Testing Coverage**: Full integration with existing test suites and interactive example
+
+## [2.4.5] - 2025-08-06 (GPT-OSS Models)
+### Added
+- Support for OpenAI GPT-OSS models on AWS Bedrock
+  - GPT-OSS-120B (120B parameter open weight model)
+  - GPT-OSS-20B (20B parameter open weight model)
+  - GPT-OSS-120B-Thinking (with reasoning tag preservation)
+  - GPT-OSS-20B-Thinking (with reasoning tag preservation)
+- `<reasoning>` tag processing for GPT-OSS thinking variants
+  - Regular GPT-OSS models automatically strip `<reasoning>` tags
+  - Thinking variants preserve `<reasoning>` tags (similar to Claude's `<think>` tags)
+- Non-streaming support for GPT-OSS models (streaming not supported by AWS Bedrock)
+- OpenAI-compatible API format with `max_completion_tokens` parameter
+
+### Technical Details
+- **Model Configuration**: All GPT-OSS models use standard messages API format
+- **API Compatibility**: Supports OpenAI-style requests with Apache 2.0 licensed models
+- **Response Processing**: Automatic reasoning tag handling based on model variant
+- **Streaming Fallback**: Automatic detection and fallback to non-streaming for unsupported models
+- **Testing Coverage**: Full integration with existing test suites and interactive example
+
 ## [2.4.4] - 2025-08-05 (Claude 4.1 Opus)
 ### Added
 - Support for Claude 4.1 Opus models

package/README.md

@@ -48,7 +48,7 @@ Bedrock Wrapper is an npm package that simplifies the integration of existing Op
     };
     ```
 
-    the `messages` variable should be in openai's role/content format  
+    the `messages` variable should be in openai's role/content format (not all models support system prompts)
     ```javascript
     messages = [
         {
@@ -98,6 +98,24 @@ Bedrock Wrapper is an npm package that simplifies the integration of existing Op
         console.log(`\n\completeResponse:\n${completeResponse}\n`); // ⇠ do stuff with the complete response
     }
 
+5. **NEW: Using the Converse API (optional)**
+    
+    You can now optionally use AWS Bedrock's Converse API instead of the Invoke API by passing `useConverseAPI: true` in the options parameter:
+    ```javascript
+    // Use the Converse API for unified request/response format across all models
+    for await (const chunk of bedrockWrapper(awsCreds, openaiChatCompletionsCreateObject, { useConverseAPI: true })) {
+        completeResponse += chunk;
+        process.stdout.write(chunk);
+    }
+    ```
+    
+    The Converse API provides:
+    - **Consistent API**: Single request/response format across all models
+    - **Simplified conversation management**: Better handling of multi-turn conversations
+    - **System prompts**: Cleaner separation of system instructions
+    - **Tool use support**: Native support for function calling (where supported)
+    - **Unified multimodal**: Consistent handling of text and image inputs
+
 ---
 
 ### Supported Models
@@ -119,6 +137,10 @@ Bedrock Wrapper is an npm package that simplifies the integration of existing Op
 | Nova-Pro                   | us.amazon.nova-pro-v1:0                      |  ✅  |
 | Nova-Lite                  | us.amazon.nova-lite-v1:0                     |  ✅  |
 | Nova-Micro                 | us.amazon.nova-micro-v1:0                    |  ❌  |
+| GPT-OSS-120B               | openai.gpt-oss-120b-1:0                      |  ❌  |
+| GPT-OSS-120B-Thinking      | openai.gpt-oss-120b-1:0                      |  ❌  |
+| GPT-OSS-20B                | openai.gpt-oss-20b-1:0                       |  ❌  |
+| GPT-OSS-20B-Thinking       | openai.gpt-oss-20b-1:0                       |  ❌  |
 | Llama-3-3-70b              | us.meta.llama3-3-70b-instruct-v1:0           |  ❌  |
 | Llama-3-2-1b               | us.meta.llama3-2-1b-instruct-v1:0            |  ❌  |
 | Llama-3-2-3b               | us.meta.llama3-2-3b-instruct-v1:0            |  ❌  |
@@ -146,7 +168,7 @@ Please modify the `bedrock_models.js` file and submit a PR 🏆 or create an Iss
 
 ### Image Support
 
-For models with image support (Claude 4 series, Claude 3.7 Sonnet, Claude 3.5 Sonnet, Claude 3 Haiku, Nova Pro, and Nova Lite), you can include images in your messages using the following format:
+For models with image support (Claude 4 series, Claude 3.7 Sonnet, Claude 3.5 Sonnet, Claude 3 Haiku, Nova Pro, and Nova Lite), you can include images in your messages using the following format (not all models support system prompts):
 
 ```javascript
 messages = [
@@ -210,6 +232,7 @@ const openaiChatCompletionsCreateObject = {
 **Model Support:**
 - ✅ **Claude models**: Fully supported (up to 8,191 sequences)
 - ✅ **Nova models**: Fully supported (up to 4 sequences)  
+- ✅ **GPT-OSS models**: Fully supported
 - ✅ **Mistral models**: Fully supported (up to 10 sequences)
 - ❌ **Llama models**: Not supported (AWS Bedrock limitation)
 
@@ -234,6 +257,44 @@ const result = await bedrockWrapper(awsCreds, {
 
 ---
 
+### 🧪 Testing
+
+The package includes comprehensive test suites to verify functionality:
+
+```bash
+# Test all models with the Both APIs (Comparison)
+npm run test
+
+# Test all models with the Invoke API
+npm run test:invoke
+
+# Test all models with the Converse API
+npm run test:converse
+
+# Test vision/multimodal capabilities with Both APIs (Comparison)
+npm run test-vision
+
+# Test vision/multimodal capabilities with Invoke API
+npm run test-vision:invoke
+
+# Test vision/multimodal capabilities with Converse API
+npm run test-vision:converse
+
+# Test stop sequences functionality with Both APIs (Comparison)
+npm run test-stop
+
+# Test stop sequences functionality with Invoke API
+npm run test-stop:invoke
+
+# Test stop sequences functionality with Converse API
+npm run test-stop:converse
+
+# Interactive testing
+npm run interactive
+```
+
+---
+
 ### 📢 P.S.
 
 In case you missed it at the beginning of this doc, for an even easier setup, use the 🔀 [Bedrock Proxy Endpoint](https://github.com/jparkerweb/bedrock-proxy-endpoint) project to spin up your own custom OpenAI server endpoint (using the standard `baseUrl`, and `apiKey` params).

package/bedrock-models.js

@@ -678,6 +678,84 @@ export const bedrock_models = [
             "inferenceConfig": {}
         }
     },
+    {
+        // ====================
+        // == GPT-OSS-120B ==
+        // ====================
+        "modelName":                     "GPT-OSS-120B",
+        "modelId":                       "openai.gpt-oss-120b-1:0",
+        // "modelId":                       "us.openai.gpt-oss-120b-1:0",
+        "vision":                        false,
+        "messages_api":                  true,
+        "system_as_separate_field":      false,
+        "display_role_names":            true,
+        "max_tokens_param_name":         "max_completion_tokens",
+        "max_supported_response_tokens": 5000,
+        "stop_sequences_param_name":     "stop_sequences",
+        "response_chunk_element":        "choices[0].delta.content",
+        "response_nonchunk_element":     "choices[0].message.content",
+        "streaming_supported":           false,
+        "special_request_schema": {}
+    },
+    {
+        // ===================
+        // == GPT-OSS-20B ==
+        // ===================
+        "modelName":                     "GPT-OSS-20B",
+        "modelId":                       "openai.gpt-oss-20b-1:0",
+        // "modelId":                       "us.openai.gpt-oss-20b-1:0",
+        "vision":                        false,
+        "messages_api":                  true,
+        "system_as_separate_field":      false,
+        "display_role_names":            true,
+        "max_tokens_param_name":         "max_completion_tokens",
+        "max_supported_response_tokens": 5000,
+        "stop_sequences_param_name":     "stop_sequences",
+        "response_chunk_element":        "choices[0].delta.content",
+        "response_nonchunk_element":     "choices[0].message.content",
+        "streaming_supported":           false,
+        "special_request_schema": {}
+    },
+    {
+        // ==============================
+        // == GPT-OSS-120B-Thinking ==
+        // ==============================
+        "modelName":                     "GPT-OSS-120B-Thinking",
+        "modelId":                       "openai.gpt-oss-120b-1:0",
+        // "modelId":                       "us.openai.gpt-oss-120b-1:0",
+        "vision":                        false,
+        "messages_api":                  true,
+        "system_as_separate_field":      false,
+        "display_role_names":            true,
+        "max_tokens_param_name":         "max_completion_tokens",
+        "max_supported_response_tokens": 5000,
+        "stop_sequences_param_name":     "stop_sequences",
+        "response_chunk_element":        "choices[0].delta.content",
+        "response_nonchunk_element":     "choices[0].message.content",
+        "streaming_supported":           false,
+        "preserve_reasoning":            true,
+        "special_request_schema": {}
+    },
+    {
+        // =============================
+        // == GPT-OSS-20B-Thinking ==
+        // =============================
+        "modelName":                     "GPT-OSS-20B-Thinking",
+        "modelId":                       "openai.gpt-oss-20b-1:0",
+        // "modelId":                       "us.openai.gpt-oss-20b-1:0",
+        "vision":                        false,
+        "messages_api":                  true,
+        "system_as_separate_field":      false,
+        "display_role_names":            true,
+        "max_tokens_param_name":         "max_completion_tokens",
+        "max_supported_response_tokens": 5000,
+        "stop_sequences_param_name":     "stop_sequences",
+        "response_chunk_element":        "choices[0].delta.content",
+        "response_nonchunk_element":     "choices[0].message.content",
+        "streaming_supported":           false,
+        "preserve_reasoning":            true,
+        "special_request_schema": {}
+    },
     {
         // ================
         // == Mistral-7b ==