### Example Configuration File

Source: https://github.com/vibheksoni/uniclaudeproxy/blob/master/README.md

This is an example configuration file. Copy it to config.json to set up your application.

```json
{
  "port": 8000,
  "log_level": "INFO",
  "providers": {
    "openai": {
      "api_key": "YOUR_OPENAI_API_KEY",
      "base_url": "https://api.openai.com/v1"
    },
    "anthropic": {
      "api_key": "YOUR_ANTHROPIC_API_KEY"
    },
    "gemini": {
      "api_key": "YOUR_GEMINI_API_KEY"
    }
  },
  "routes": [
    {
      "path": "/v1/chat/completions",
      "provider": "openai",
      "converter": "anthropic_to_openai"
    },
    {
      "path": "/v1/completions",
      "provider": "anthropic",
      "converter": "openai_to_anthropic"
    },
    {
      "path": "/v1/gemini/chat",
      "provider": "gemini",
      "converter": "anthropic_to_gemini"
    }
  ]
}
```

--------------------------------

### Clone Repository and Install Dependencies

Source: https://github.com/vibheksoni/uniclaudeproxy/blob/master/README.md

Clone the UniClaudeProxy repository and install the required Python packages.

```bash
git clone https://github.com/vibheksoni/UniClaudeProxy.git
cd UniClaudeProxy
pip install -r requirements.txt
```

--------------------------------

### Start UniClaudeProxy on Linux/macOS

Source: https://github.com/vibheksoni/uniclaudeproxy/blob/master/README.md

Make the Run.sh script executable and then run it to start the proxy service on Linux or macOS.

```bash
chmod +x Run.sh
./Run.sh
```

--------------------------------

### Start UniClaudeProxy on Windows

Source: https://github.com/vibheksoni/uniclaudeproxy/blob/master/README.md

Execute the Run.bat script to start the proxy service on a Windows environment.

```bash
Run.bat
```

--------------------------------

### Create and Edit Configuration File

Source: https://github.com/vibheksoni/uniclaudeproxy/blob/master/README.md

Copy the example configuration file and edit it with your specific API keys and model mappings.

```bash
cp config.example.json config.json
```

--------------------------------

### Start UniClaudeProxy Directly

Source: https://github.com/vibheksoni/uniclaudeproxy/blob/master/README.md

Run the UniClaudeProxy service directly using uvicorn, specifying the host and port.

```bash
python -m uvicorn app.main:app --host 127.0.0.1 --port 9223
```

--------------------------------

### Launch Claude Code with Profile

Source: https://github.com/vibheksoni/uniclaudeproxy/blob/master/README.md

Start Claude Code using the configured profile to direct its requests through the UniClaudeProxy.

```bash
claude --profile cc-proxy
```

--------------------------------

### ReAct XML Observation Example

Source: https://github.com/vibheksoni/uniclaudeproxy/blob/master/README.md

Demonstrates how tool call results are received in XML format when using the ReAct XML tool calling mechanism.

```xml
<observation>
total 42
drwxr-xr-x  5 user user 4096 Feb  9 12:00 .
...
</observation>
```

--------------------------------

### System Prompt Replacement Example

Source: https://github.com/vibheksoni/uniclaudeproxy/blob/master/README.md

Use `system_replacements` to modify system prompts for models that reject identity claims. Replacements are applied universally before provider dispatch.

```json
"system_replacements": {
  "You are Claude Code, Anthropic's official CLI for Claude.": "You are an advanced AI coding assistant integrated into a CLI tool.",
  "Claude Code": "the coding assistant"
}
```

--------------------------------

### ReAct XML Tool Call Example

Source: https://github.com/vibheksoni/uniclaudeproxy/blob/master/README.md

Shows how models without native function calling can use XML to describe tool calls. This is used when 'use_react' is enabled in the model configuration.

```xml
<tool_call>
<name>Bash</name>
<parameters>
{"command": "ls -la"}
</parameters>
</tool_call>
```

--------------------------------

### Configure Google Gemini Provider

Source: https://github.com/vibheksoni/uniclaudeproxy/blob/master/README.md

Configure UniClaudeProxy to connect to the Google Gemini API. This setup supports Gemini's advanced features like function calling.

```json
{
  "provider_type": "gemini",
  "api_key": "your-gemini-key",
  "base_url": "https://generativelanguage.googleapis.com/v1beta",
  "headers": {
    "x-goog-api-key": "your-gemini-key"
  },
  "models": {
    "gemini-2.5-pro-preview-06-05": {
      "name": "Gemini 2.5 Pro"
    }
  }
}
```

--------------------------------

### UniClaudeProxy Configuration File Structure

Source: https://github.com/vibheksoni/uniclaudeproxy/blob/master/README.md

The main configuration file for UniClaudeProxy, defining server, model, and provider settings.

```json
{
  "server": { "host": "127.0.0.1", "port": 9223, "local_only": true },
  "models": {
    "<anthropic-model-name>": "<provider-name>/<model-id>"
  },
  "providers": {
    "<provider-name>": {
      "provider_type": "openai | gemini | claude",
      "api_key": "your-api-key",
      "base_url": "https://api.example.com",
      "headers": {},
      "models": {
        "<model-id>": { ... }
      }
    }
  }
}
```

--------------------------------

### Configure OpenAI-Compatible Provider

Source: https://github.com/vibheksoni/uniclaudeproxy/blob/master/README.md

Set up UniClaudeProxy to use an OpenAI-compatible API endpoint. This is useful for various services like DeepSeek, Ollama, vLLM, and cloud inference providers.

```json
{
  "provider_type": "openai",
  "api_key": "your-key",
  "base_url": "https://api.deepseek.com",
  "models": {
    "deepseek-chat": {
      "name": "DeepSeek V3",
      "max_output_tokens": 8192
    }
  }
}
```

--------------------------------

### Check Tool Calls in Debug Log

Source: https://github.com/vibheksoni/uniclaudeproxy/blob/master/README.md

Use this command to verify if tool calls are working correctly by searching for 'REACT STREAM' or 'tool_use=True' in the debug log and showing the last 10 lines.

```bash
# Check tool calls working
grep "REACT STREAM\|tool_use=True" debug.log | tail -10
```

--------------------------------

### UniClaudeProxy Request Flow Diagram

Source: https://github.com/vibheksoni/uniclaudeproxy/blob/master/README.md

Illustrates the request flow from Claude Code CLI through UniClaudeProxy to any LLM provider, including the conversion and streaming process.

```text
Claude Code CLI  -->  UniClaudeProxy (localhost:9223)  -->  Any LLM Provider
(Anthropic fmt)       Route + Convert + Stream              (OpenAI/Gemini/etc)
     ^                        |
     +---- Anthropic SSE <----+

```

--------------------------------

### Enable Local-Only Mode

Source: https://github.com/vibheksoni/uniclaudeproxy/blob/master/README.md

Configure UniClaudeProxy to only accept connections from localhost for enhanced security. This setting is enabled by default.

```json
{
  "server": {
    "host": "127.0.0.1",
    "port": 9223,
    "local_only": true
  }
}
```

--------------------------------

### Configure Anthropic Passthrough Provider

Source: https://github.com/vibheksoni/uniclaudeproxy/blob/master/README.md

Set up UniClaudeProxy to forward requests directly to an Anthropic-compatible API. This configuration bypasses any conversion, forwarding the raw request body.

```json
{
  "provider_type": "claude",
  "api_key": "your-key",
  "base_url": "https://api.anthropic.com/v1",
  "headers": {
    "anthropic-version": "2023-06-01"
  },
  "models": {
    "claude-sonnet-4-5-20250929": {
      "name": "Claude Sonnet 4.5"
    }
  }
}
```

--------------------------------

### UniClaudeProxy Architecture Diagram

Source: https://github.com/vibheksoni/uniclaudeproxy/blob/master/README.md

This diagram illustrates the flow of requests from Claude Code CLI through UniClaudeProxy to various LLM providers, including Anthropic SSE.

```text
Claude Code CLI  -->  UniClaudeProxy (localhost:9223)  -->  Any LLM Provider
     ^                        |
     +---- Anthropic SSE <----+
```

--------------------------------

### Configure Claude Code Profile

Source: https://github.com/vibheksoni/uniclaudeproxy/blob/master/README.md

Set up a Claude Code profile in settings.json to route traffic through UniClaudeProxy. Ensure ANTHROPIC_BASE_URL is set to the proxy's address.

```json
{
  "profiles": {
    "cc-proxy": {
      "env": {
        "ANTHROPIC_AUTH_TOKEN": "",
        "ANTHROPIC_BASE_URL": "http://127.0.0.1:9223",
        "CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC": "1",
        "API_TIMEOUT_MS": "3000000",
        "CLAUDE_CODE_MAX_OUTPUT_TOKENS": "50000",
        "CLAUDE_BASH_NO_LOGIN": "1"
      },
      "permissions": {
        "allow": [],
        "deny": []
      }
    }
  }
}
```

--------------------------------

### Check Route Resolution in Debug Log

Source: https://github.com/vibheksoni/uniclaudeproxy/blob/master/README.md

Use this command to check how routes are being resolved by filtering the debug log for 'Request: model=' and showing the last 5 lines.

```bash
# Check route resolution
grep "Request: model=" debug.log | tail -5
```

--------------------------------

### Check for Errors in Debug Log

Source: https://github.com/vibheksoni/uniclaudeproxy/blob/master/README.md

Use this command to quickly identify errors by filtering the debug log for 'ERROR' or 'Traceback' and displaying the last 10 lines.

```bash
# Check for errors
grep "ERROR\|Traceback" debug.log | tail -10
```

--------------------------------

### See Outgoing Requests in Debug Log

Source: https://github.com/vibheksoni/uniclaudeproxy/blob/master/README.md

Use this command to inspect what is being sent upstream by searching for 'OUTGOING REQUEST' in the debug log and showing the last 5 lines.

```bash
# See what's being sent upstream
grep "OUTGOING REQUEST" debug.log | tail -5
```

=== COMPLETE CONTENT === This response contains all available snippets from this library. No additional content exists. Do not make further requests.