Aurai Advisor

Provides remote guidance for local AI with multi-provider support, persistent history, and GUI configuration.

python

GitHub Stars

python

Language

4 months ago

First Indexed

2 months ago

Catalog Refreshed

Documentation & install

Readme and setup notes from the catalogue, plus a client-ready config you can copy for your MCP host.

Installation

Add the following to your MCP client configuration file.

Configuration

View docs

{
  "mcpServers": {
    "lzmw-mcp-aurai-server": {
      "command": "D:\\mcp-aurai-server\\venv\\Scripts\\python.exe",
      "args": [
        "-m",
        "mcp_aurai.server"
      ],
      "env": {
        "AURAI_MODEL": "deepseek-v3-1-250821",
        "AURAI_API_KEY": "YOUR_API_KEY",
        "AURAI_BASE_URL": "https://www.chatgtp.cn/v1",
        "AURAI_PROVIDER": "custom",
        "AURAI_LOG_LEVEL": "INFO",
        "AURAI_MAX_TOKENS": "100000",
        "AURAI_MAX_HISTORY": "50",
        "AURAI_TEMPERATURE": "1.0",
        "AURAI_HISTORY_PATH": "~/.mcp-aurai/history.json",
        "AURAI_MAX_ITERATIONS": "10",
        "AURAI_ENABLE_PERSISTENCE": "true"
      }
    }
  }
}

You use the MCP Aurai Advisor to give your local AI access to remote, expert guidance. It coordinates several AI providers, manages conversation history, and lets you run a local MCP server that can be guided by an upstream advisor for solving programming and design problems.

How to use

After you configure the MCP, restart your Claude Code session. In a chat, simply describe the programming issue you are facing. If needed, paste relevant code or error messages. The advisor will decide whether to invoke the consult_aurai tool to get expert guidance from the remote advisor, and it will manage the conversation context automatically to keep interactions focused on your current question.

How to install

Prerequisites: you need Python and Git (for cloning or installing dependencies). You will also use a Python virtual environment to isolate this MCP server.

Step 1 — Prepare the environment and install the package:

# 1. Create or navigate to the project directory
cd mcp-aurai-server

# 2. Create a virtual environment
python -m venv venv

# 3. Activate the virtual environment
# Windows
venv\Scripts\activate
# macOS/Linux
# source venv/bin/activate

# 4. Install dependencies
pip install -e ".[all-dev]"

# 5. Run tests to verify installation
pytest tests/ -v
# Expected: ========== 27 passed ==========

Configuration and run approach

You can configure the MCP in two ways. The recommended way is to use the configuration tool to generate a complete setup. The manual method is also supported if you want to customize the exact startup command and environment.

Option A — Configuration tool (recommended):

python tools\control_center.py

Option B — Manual configuration using the explicit MCP run command shown here. This configures the local advisor to run inside your Python virtual environment and uses a custom provider.

claude mcp add --scope user --transport stdio aurai-advisor \
  --env AURAI_API_KEY="your-api-key" \
  --env AURAI_PROVIDER="custom" \
  --env AURAI_BASE_URL="https://www.chatgtp.cn/v1" \
  --env AURAI_MODEL="deepseek-v3-1-250821" \
  -- "D:\\mcp-aurai-server\\venv\\Scripts\\python.exe" "-m" "mcp_aurai.server"

Additional configuration notes

Environment variables shown are required for configuring the remote advisor and controlling behavior. You can adjust optional values to influence iteration limits, history persistence, and model selection.

Available tools

consult_aurai

Primary MCP tool to request guidance from the upper-level AI advisor for programming problems, including problem type, error messages, and code context.

sync_context

Upload documentation or code context to the advisor to ensure complete understanding without truncation.

report_progress

Report the progress after following the advisor’s guidance to get next steps.

get_status

Check the current conversation status, iteration count, and configuration state.

Built by

VeilStrat

AI signals for GTM teams