Added bedrock support

Author: vishal veerareddy

.env.example

@@ -119,6 +119,92 @@ OLLAMA_MAX_TOOLS_FOR_ROUTING=3
 # Optional API key (for secured LM Studio servers)
 # LMSTUDIO_API_KEY=your-optional-api-key
 
+# ==============================================================================
+# AWS Bedrock Configuration (Claude, Titan, Llama, Jurassic, etc.)
+# ==============================================================================
+
+# Bedrock API Key (Bearer token) - REQUIRED
+# Generate from AWS Console → Bedrock → API Keys
+# See: https://docs.aws.amazon.com/bedrock/latest/userguide/api-keys-generate.html
+# AWS_BEDROCK_API_KEY=your-bedrock-api-key-here
+
+# AWS region (default: us-east-1)
+# Available regions: us-east-1, us-west-2, us-east-2, ap-southeast-1, ap-northeast-1, eu-central-1
+# AWS_BEDROCK_REGION=us-east-1
+
+# Bedrock model ID to use (via Converse API - unified format for all text models)
+#
+# ===== CLAUDE MODELS (BEST FOR TOOL CALLING) =====
+# Claude 4.5 models (latest - requires inference profiles):
+#   - us.anthropic.claude-sonnet-4-5-20250929-v1:0 (regional US, best for tool calling)
+#   - us.anthropic.claude-haiku-4-5-20251001-v1:0 (fast, efficient)
+#   - global.anthropic.claude-sonnet-4-5-20250929-v1:0 (cross-region)
+# Claude 3.x models:
+#   - anthropic.claude-3-5-sonnet-20241022-v2:0 (excellent for tool calling)
+#   - anthropic.claude-3-opus-20240229-v1:0 (most capable)
+#   - anthropic.claude-3-haiku-20240307-v1:0 (fast, cheap)
+#
+# ===== DEEPSEEK MODELS (NEW - 2025) =====
+#   - us.deepseek.r1-v1:0 (DeepSeek R1 - reasoning model)
+#   - deepseek.r1-v1:0 (direct model ID)
+#
+# ===== QWEN MODELS (ALIBABA - NEW 2025) =====
+#   - qwen.qwen3-235b-a22b-2507-v1:0 (largest, 235B parameters)
+#   - qwen.qwen3-32b-v1:0 (balanced, 32B parameters)
+#   - qwen.qwen3-coder-480b-a35b-v1:0 (coding specialist, 480B)
+#   - qwen.qwen3-coder-30b-a3b-v1:0 (coding, smaller)
+#
+# ===== OPENAI OPEN-WEIGHT MODELS (NEW - 2025) =====
+#   - openai.gpt-oss-120b-1:0 (120B parameters, open-weight)
+#   - openai.gpt-oss-20b-1:0 (20B parameters, efficient)
+#
+# ===== GOOGLE GEMMA MODELS (OPEN-WEIGHT) =====
+#   - google.gemma-3-27b (27B parameters)
+#   - google.gemma-3-12b (12B parameters)
+#   - google.gemma-3-4b (4B parameters, efficient)
+#
+# ===== MINIMAX MODELS (NEW - 2025) =====
+#   - minimax.m2-v1:0 (MiniMax M2)
+#
+# ===== AMAZON MODELS =====
+# Nova (multimodal):
+#   - us.amazon.nova-pro-v1:0 (best quality, multimodal)
+#   - us.amazon.nova-lite-v1:0 (fast, cost-effective)
+#   - us.amazon.nova-micro-v1:0 (ultra-fast, text-only)
+# Titan:
+#   - amazon.titan-text-premier-v1:0 (largest)
+#   - amazon.titan-text-express-v1 (fast)
+#   - amazon.titan-text-lite-v1 (cheapest)
+#
+# ===== META LLAMA MODELS =====
+#   - meta.llama3-1-70b-instruct-v1:0 (most capable)
+#   - meta.llama3-1-8b-instruct-v1:0 (fast, efficient)
+#
+# ===== MISTRAL MODELS =====
+#   - mistral.mistral-large-2407-v1:0 (largest)
+#   - mistral.mistral-small-2402-v1:0 (efficient)
+#   - mistral.mixtral-8x7b-instruct-v0:1 (mixture of experts)
+#
+# ===== COHERE COMMAND MODELS =====
+#   - cohere.command-r-plus-v1:0 (best for RAG)
+#   - cohere.command-r-v1:0 (balanced)
+#
+# ===== AI21 MODELS =====
+#   - ai21.jamba-1-5-large-v1:0 (hybrid architecture)
+#   - ai21.jamba-1-5-mini-v1:0 (fast)
+#
+# ===== OTHER PROVIDERS (CHECK AWS CONSOLE FOR AVAILABILITY) =====
+# Writer AI, Kimi AI, Luma AI, TwelveLabs models may also be available
+#
+# Default (Claude Sonnet 4.5 with tool calling):
+# AWS_BEDROCK_MODEL_ID=us.anthropic.claude-sonnet-4-5-20250929-v1:0
+
+# Cost comparison (per 1M tokens):
+# Claude 3.5 Sonnet on Bedrock: $3.00 input / $15.00 output
+# Claude 3 Opus on Bedrock: $15.00 input / $75.00 output
+# Titan Text Express: $0.20 input / $0.60 output
+# Llama 3 70B: $0.99 input / $0.99 output
+
 # ==============================================================================
 # Server Configuration
 # ==============================================================================

README.md

@@ -5,17 +5,18 @@
 [![License: Apache 2.0](https://img.shields.io/badge/license-Apache%202.0-blue.svg)](LICENSE "Apache 2.0 License - Open Source Claude Code Alternative")
 [![Ask DeepWiki](https://deepwiki.com/badge.svg)](https://deepwiki.com/vishalveerareddy123/Lynkr "Lynkr Documentation on DeepWiki")
 [![Databricks Supported](https://img.shields.io/badge/Databricks-Supported-orange)](https://www.databricks.com/ "Databricks Claude Integration")
+[![AWS Bedrock](https://img.shields.io/badge/AWS%20Bedrock-100%2B%20Models-FF9900)](https://aws.amazon.com/bedrock/ "AWS Bedrock - 100+ Models")
 [![OpenAI Compatible](https://img.shields.io/badge/OpenAI-Compatible-412991)](https://openai.com/ "OpenAI GPT Integration")
 [![Ollama Compatible](https://img.shields.io/badge/Ollama-Compatible-brightgreen)](https://ollama.ai/ "Local Ollama Model Support")
 [![llama.cpp Compatible](https://img.shields.io/badge/llama.cpp-Compatible-blue)](https://github.com/ggerganov/llama.cpp "llama.cpp GGUF Model Support")
 [![IndexNow Enabled](https://img.shields.io/badge/IndexNow-Enabled-success?style=flat-square)](https://www.indexnow.org/ "SEO Optimized with IndexNow")
 [![DevHunt](https://img.shields.io/badge/DevHunt-Lynkr-orange)](https://devhunt.org/tool/lynkr "Lynkr on DevHunt")
 
-> ** Claude Code proxy server supporting Databricks, OpenRouter, Ollama & Azure. Features MCP integration, prompt caching & 60-80% token optimization savings.**
+> ** Claude Code proxy server supporting Databricks, AWS Bedrock (100+ models), OpenRouter, Ollama & Azure. Features MCP integration, prompt caching & 60-80% token optimization savings.**
 
 ## 🔖 Keywords
 
-`claude-code` `claude-proxy` `anthropic-api` `databricks-llm` `openrouter-integration` `ollama-local` `llama-cpp` `azure-openai` `azure-anthropic` `mcp-server` `prompt-caching` `token-optimization` `ai-coding-assistant` `llm-proxy` `self-hosted-ai` `git-automation` `code-generation` `developer-tools` `ci-cd-automation` `llm-gateway` `cost-reduction` `multi-provider-llm`
+`claude-code` `claude-proxy` `anthropic-api` `databricks-llm` `aws-bedrock` `bedrock-models` `deepseek-r1` `qwen3-coder` `openrouter-integration` `ollama-local` `llama-cpp` `azure-openai` `azure-anthropic` `mcp-server` `prompt-caching` `token-optimization` `ai-coding-assistant` `llm-proxy` `self-hosted-ai` `git-automation` `code-generation` `developer-tools` `ci-cd-automation` `llm-gateway` `cost-reduction` `multi-provider-llm`
 
 ---
 
@@ -68,7 +69,7 @@ Claude Code CLI is locked to Anthropic's API, limiting your choice of LLM provid
 ### The Solution
 Lynkr is a **production-ready proxy server** that unlocks Claude Code CLI's full potential:
 
-- ✅ **Any LLM Provider** - [Databricks, OpenRouter (100+ models), Ollama (local), Azure, OpenAI, llama.cpp](#supported-ai-model-providers-databricks-openrouter-ollama-azure-llamacpp)
+- ✅ **Any LLM Provider** - [Databricks, AWS Bedrock (100+ models), OpenRouter (100+ models), Ollama (local), Azure, OpenAI, llama.cpp](#supported-ai-model-providers-databricks-aws-bedrock-openrouter-ollama-azure-llamacpp)
 - ✅ **60-80% Cost Reduction** - Built-in [token optimization](#token-optimization-implementation) (5 optimization phases implemented)
 - ✅ **Zero Code Changes** - [Drop-in replacement](#connecting-claude-code-cli) for Anthropic backend
 - ✅ **Local & Offline** - Run Claude Code with [Ollama](#using-ollama-models) or [llama.cpp](#using-llamacpp-with-lynkr) (no internet required)
@@ -94,11 +95,17 @@ npm install -g lynkr
 
 ### 2️⃣ Configure Your Provider
 ```bash
-# Option A: Use local Ollama (free, offline)
+# Option A: Use AWS Bedrock (100+ models) 🆕
+export MODEL_PROVIDER=bedrock
+export AWS_BEDROCK_API_KEY=your-bearer-token
+export AWS_BEDROCK_REGION=us-east-2
+export AWS_BEDROCK_MODEL_ID=us.anthropic.claude-sonnet-4-5-20250929-v1:0
+
+# Option B: Use local Ollama (free, offline)
 export MODEL_PROVIDER=ollama
 export OLLAMA_MODEL=llama3.1:8b
 
-# Option B: Use Databricks (production)
+# Option C: Use Databricks (production)
 export MODEL_PROVIDER=databricks
 export DATABRICKS_API_BASE=https://your-workspace.databricks.net
 export DATABRICKS_API_KEY=your-api-key
@@ -160,7 +167,7 @@ Further documentation and usage notes are available on [DeepWiki](https://deepwi
 
 ---
 
-## Supported AI Model Providers (Databricks, OpenRouter, Ollama, Azure, llama.cpp)
+## Supported AI Model Providers (Databricks, AWS Bedrock, OpenRouter, Ollama, Azure, llama.cpp)
 
 Lynkr supports multiple AI model providers, giving you flexibility in choosing the right model for your needs:
 
@@ -169,6 +176,7 @@ Lynkr supports multiple AI model providers, giving you flexibility in choosing t
 | Provider | Configuration | Models Available | Best For |
 |----------|--------------|------------------|----------|
 | **Databricks** (Default) | `MODEL_PROVIDER=databricks` | Claude Sonnet 4.5, Claude Opus 4.5 | Production use, enterprise deployment |
+| **AWS Bedrock** 🆕 | `MODEL_PROVIDER=bedrock` | 100+ models (Claude, DeepSeek R1, Qwen3, Nova, Titan, Llama, Mistral, etc.) | AWS ecosystem, multi-model flexibility, Claude + alternatives |
 | **OpenAI** | `MODEL_PROVIDER=openai` | GPT-5, GPT-5.2, GPT-4o, GPT-4o-mini, GPT-4-turbo, o1, o1-mini | Direct OpenAI API access |
 | **Azure OpenAI** | `MODEL_PROVIDER=azure-openai` | GPT-5, GPT-5.2,GPT-4o, GPT-4o-mini, GPT-5, o1, o3, Kimi-K2 | Azure integration, Microsoft ecosystem |
 | **Azure Anthropic** | `MODEL_PROVIDER=azure-anthropic` | Claude Sonnet 4.5, Claude Opus 4.5 | Azure-hosted Claude models |
@@ -204,6 +212,44 @@ When using `MODEL_PROVIDER=azure-openai`, you can deploy any of the models in az
 
 **Note**: Azure OpenAI deployment names are configurable via `AZURE_OPENAI_DEPLOYMENT` environment variable.
 
+### **AWS Bedrock Model Catalog (100+ Models)**
+
+When using `MODEL_PROVIDER=bedrock`, you have access to **nearly 100 models** via AWS Bedrock's unified Converse API:
+
+#### **🆕 NEW Models (2025-2026)**
+- **DeepSeek R1** - `us.deepseek.r1-v1:0` - Reasoning model (o1-style)
+- **Qwen3** - `qwen.qwen3-235b-*`, `qwen.qwen3-coder-480b-*` - Up to 480B parameters!
+- **OpenAI GPT-OSS** - `openai.gpt-oss-120b-1:0` - Open-weight GPT models
+- **Google Gemma 3** - `google.gemma-3-27b` - Open-weight from Google
+- **MiniMax M2** - `minimax.m2-v1:0` - Chinese AI company
+
+#### **Claude Models (Best for Tool Calling)**
+- **Claude 4.5** - `us.anthropic.claude-sonnet-4-5-*` - Best for coding with tools
+- **Claude 3.5** - `anthropic.claude-3-5-sonnet-*` - Excellent tool calling
+- **Claude 3 Haiku** - `anthropic.claude-3-haiku-*` - Fast and cost-effective
+
+#### **Amazon Models**
+- **Nova** - `us.amazon.nova-pro-v1:0` - Multimodal, 300K context
+- **Titan** - `amazon.titan-text-express-v1` - General purpose
+
+#### **Other Major Models**
+- **Meta Llama** - `meta.llama3-1-70b-*` - Open-source Llama 3.1
+- **Mistral** - `mistral.mistral-large-*` - Coding, multilingual
+- **Cohere** - `cohere.command-r-plus-v1:0` - RAG, search
+- **AI21 Jamba** - `ai21.jamba-1-5-large-v1:0` - 256K context
+
+#### **Quick Setup**
+```bash
+export MODEL_PROVIDER=bedrock
+export AWS_BEDROCK_API_KEY=your-bearer-token  # Get from AWS Console → Bedrock → API Keys
+export AWS_BEDROCK_REGION=us-east-2
+export AWS_BEDROCK_MODEL_ID=us.anthropic.claude-sonnet-4-5-20250929-v1:0
+```
+
+📖 **Full Documentation**: See [BEDROCK_MODELS.md](BEDROCK_MODELS.md) for complete model catalog, pricing, capabilities, and use cases.
+
+⚠️ **Tool Calling Note**: Only **Claude models** support tool calling on Bedrock. Other models work via Converse API but won't use Read/Write/Bash tools.
+
 ### **Ollama Model Recommendations**
 
 For tool calling support (required for Claude Code CLI functionality):
@@ -241,16 +287,19 @@ FALLBACK_PROVIDER=databricks  # or azure-openai, openrouter, azure-anthropic
 
 ### **Provider Comparison**
 
-| Feature | Databricks | OpenAI | Azure OpenAI | Azure Anthropic | OpenRouter | Ollama | llama.cpp |
-|---------|-----------|--------|--------------|-----------------|------------|--------|-----------|
-| **Setup Complexity** | Medium | Easy | Medium | Medium | Easy | Easy | Medium |
-| **Cost** | $$$ | $$ | $$ | $$$ | $ | Free | Free |
-| **Latency** | Low | Low | Low | Low | Medium | Very Low | Very Low |
-| **Tool Calling** | Excellent | Excellent | Excellent | Excellent | Good | Fair | Good |
-| **Context Length** | 200K | 128K | 128K | 200K | Varies | 32K-128K | Model-dependent |
-| **Streaming** | Yes | Yes | Yes | Yes | Yes | Yes | Yes |
-| **Privacy** | Enterprise | Third-party | Enterprise | Enterprise | Third-party | Local | Local |
-| **Offline** | No | No | No | No | No | Yes | Yes |
+| Feature | Databricks | AWS Bedrock | OpenAI | Azure OpenAI | Azure Anthropic | OpenRouter | Ollama | llama.cpp |
+|---------|-----------|-------------|--------|--------------|-----------------|------------|--------|-----------|
+| **Setup Complexity** | Medium | Easy | Easy | Medium | Medium | Easy | Easy | Medium |
+| **Cost** | $$$ | $$ | $$ | $$ | $$$ | $ | Free | Free |
+| **Latency** | Low | Low | Low | Low | Low | Medium | Very Low | Very Low |
+| **Model Variety** | 2 | 100+ | 10+ | 10+ | 2 | 100+ | 50+ | Unlimited |
+| **Tool Calling** | Excellent | Excellent* | Excellent | Excellent | Excellent | Good | Fair | Good |
+| **Context Length** | 200K | Up to 300K | 128K | 128K | 200K | Varies | 32K-128K | Model-dependent |
+| **Streaming** | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes |
+| **Privacy** | Enterprise | Enterprise | Third-party | Enterprise | Enterprise | Third-party | Local | Local |
+| **Offline** | No | No | No | No | No | No | Yes | Yes |
+
+_* Tool calling only supported by Claude models on Bedrock_
 
 ---
 

docs/index.md

@@ -6,7 +6,7 @@
 
 # Lynkr - Production-Ready Claude Code Proxy with Multi-Provider Support, MCP Integration & Token Optimization
 
-#### Lynkr is an open-source, production-ready Claude Code proxy that enables the Claude Code CLI to work with any LLM provider (Databricks, OpenRouter, Ollama, Azure, OpenAI, llama.cpp) without losing Anthropic backend features. It features MCP server orchestration, Git workflows, repo intelligence, workspace tools, prompt caching, and 60-80% token optimization for cost-effective LLM-powered development.
+#### Lynkr is an open-source, production-ready Claude Code proxy that enables the Claude Code CLI to work with any LLM provider (Databricks, AWS Bedrock (100+ models), OpenRouter, Ollama, Azure, OpenAI, llama.cpp) without losing Anthropic backend features. It features MCP server orchestration, Git workflows, repo intelligence, workspace tools, prompt caching, and 60-80% token optimization for cost-effective LLM-powered development.
 <!--
 SEO Keywords:
 Databricks, Claude Code, Anthropic, Azure Anthropic,
@@ -17,7 +17,7 @@ prompt caching, Node.js
 
 ## 🔖 Keywords
 
-`claude-code` `claude-proxy` `anthropic-api` `databricks-llm` `openrouter-integration` `ollama-local` `llama-cpp` `azure-openai` `azure-anthropic` `mcp-server` `prompt-caching` `token-optimization` `ai-coding-assistant` `llm-proxy` `self-hosted-ai` `git-automation` `code-generation` `developer-tools` `ci-cd-automation` `llm-gateway` `cost-reduction` `multi-provider-llm`
+`claude-code` `claude-proxy` `anthropic-api` `databricks-llm` `aws-bedrock` `bedrock-models` `deepseek-r1` `qwen3` `openrouter-integration` `ollama-local` `llama-cpp` `azure-openai` `azure-anthropic` `mcp-server` `prompt-caching` `token-optimization` `ai-coding-assistant` `llm-proxy` `self-hosted-ai` `git-automation` `code-generation` `developer-tools` `ci-cd-automation` `llm-gateway` `cost-reduction` `multi-provider-llm`
 
 ---
 
@@ -32,7 +32,7 @@ prompt caching, Node.js
 
 # 🚀 What is Lynkr?
 
-**Lynkr** is an open-source **Claude Code-compatible backend proxy** that lets you run the **Claude Code CLI** and Claude-style tools **directly against [Databricks, Azure, OpenRouter, Ollama, and llama.cpp](#-configuration-guide-for-multi-provider-support-databricks-azure-openrouter-ollama-llamacpp)** instead of the default Anthropic cloud.
+**Lynkr** is an open-source **Claude Code-compatible backend proxy** that lets you run the **Claude Code CLI** and Claude-style tools **directly against [Databricks, AWS Bedrock, Azure, OpenRouter, Ollama, and llama.cpp](#-configuration-guide-for-multi-provider-support-databricks-aws-bedrock-azure-openrouter-ollama-llamacpp)** instead of the default Anthropic cloud.
 
 It enables full repo-aware LLM workflows:
 
@@ -58,6 +58,9 @@ Emulates Anthropic’s backend so the **Claude Code CLI works without modificati
 ### ✔ Works with Databricks LLM Serving  
 Supports **Databricks-hosted Claude Sonnet / Haiku models**, or any LLM served from Databricks.
 
+### ✔ Supports AWS Bedrock (100+ models) 🆕
+Access Claude 4.5, DeepSeek R1, Qwen3 (480B), OpenAI GPT-OSS, Google Gemma, Nova, Titan, Llama, Mistral, and more through AWS's unified Converse API.
+
 ### ✔ Supports Azure Anthropic models
 Route Claude Code requests into Azure's `/anthropic/v1/messages` endpoint.
 
@@ -137,7 +140,7 @@ Claude Code CLI
 ↓
 Lynkr Proxy
 ↓
-Databricks / Azure Anthropic / OpenRouter / Ollama / llama.cpp / MCP / Tools
+Databricks / AWS Bedrock / Azure Anthropic / OpenRouter / Ollama / llama.cpp / MCP / Tools
 
 ```
 
@@ -265,7 +268,7 @@ npm start
 
 ---
 
-# 🔧 Configuration Guide for Multi-Provider Support (Databricks, Azure, OpenRouter, Ollama, llama.cpp)
+# 🔧 Configuration Guide for Multi-Provider Support (Databricks, AWS Bedrock, Azure, OpenRouter, Ollama, llama.cpp)
 
 ## Databricks Setup
 
@@ -299,6 +302,47 @@ AZURE_OPENAI_DEPLOYMENT=gpt-4o
 PORT=8080
 ```
 
+## AWS Bedrock Setup (🆕 NEW - 100+ Models)
+
+**What is AWS Bedrock?**
+
+AWS Bedrock provides serverless access to **nearly 100 foundation models** from leading AI companies through a unified Converse API. Benefits:
+- ✅ **Massive model variety** – Claude, DeepSeek R1, Qwen3, OpenAI GPT-OSS, Google Gemma, Nova, Titan, Llama, Mistral, Cohere, and more
+- ✅ **Enterprise-grade** – AWS infrastructure, security, and compliance
+- ✅ **Pay-per-use** – No monthly fees, pay only for tokens used
+- ✅ **Claude tool calling** – Full Read/Write/Bash tool support with Claude models
+- ✅ **Multi-cloud flexibility** – Not locked into single provider
+
+**Configuration:**
+
+```env
+MODEL_PROVIDER=bedrock
+AWS_BEDROCK_API_KEY=your-bearer-token           # Get from AWS Console → Bedrock → API Keys
+AWS_BEDROCK_REGION=us-east-2                    # us-east-1, us-west-2, us-east-2, etc.
+AWS_BEDROCK_MODEL_ID=us.anthropic.claude-sonnet-4-5-20250929-v1:0
+PORT=8080
+WORKSPACE_ROOT=/path/to/your/repo
+```
+
+**Popular Models:**
+- `us.anthropic.claude-sonnet-4-5-20250929-v1:0` – Best for coding with tool calling
+- `us.deepseek.r1-v1:0` – Reasoning model (o1-style)
+- `qwen.qwen3-coder-480b-a35b-v1:0` – 480B parameter coding specialist
+- `openai.gpt-oss-120b-1:0` – Open-weight GPT model
+- `us.amazon.nova-pro-v1:0` – Multimodal, 300K context
+- `meta.llama3-1-70b-instruct-v1:0` – Open-source Llama 3.1
+- `mistral.mistral-large-2407-v1:0` – Multilingual, coding
+
+📖 See [BEDROCK_MODELS.md](https://github.com/vishalveerareddy123/Lynkr/blob/main/BEDROCK_MODELS.md) for complete model catalog (100+ models), pricing, capabilities, and use cases.
+
+**Getting Started:**
+1. Visit [AWS Bedrock Console](https://console.aws.amazon.com/bedrock/)
+2. Navigate to API Keys section
+3. Generate a new API key (Bearer token)
+4. Configure Lynkr as shown above
+5. Start using any of the 100+ available models!
+
+⚠️ **Tool Calling Note**: Only Claude models support tool calling (Read, Write, Bash) on Bedrock. Other models work via Converse API but won't execute tools.
 
 ## OpenRouter Setup