Ai_agents

Ai_agents is a multi-agent system harnessed through MCP that orchestrates several domain-specific agents: a Scraper Agent for web research, a Podcast Agent for dynamic podcast generation using text-to-speech, a Stock Agent for real-time and historical market data, a GitHub Agent for repository management and documentation, a Notion Copilot for research and Notion automation, and a RAG tool for retrieve-and-present results. The server uses LangGraph and LangChain under MCP to route tasks to the appropriate agent and combine results into a coherent output. To use it, install the MCP runtime with uvx as described in the installation steps, run the Ai_agents MCP server, and then interact with the system via its MCP-compatible interface or through any provided UI endpoints. Tools like web scraping (Tavily/Firecrawl), TTS (ElevenLabs), stock data providers (Alpha Vantage, NSE, MoneyControl), GitHub API, and Notion APIs are wired into the agents to perform specialized actions. You can request comprehensive research, automated content generation, or repository maintenance, and the MCP layer will manage tool invocation and result synthesis across the agents.

Prerequisites:

• Python 3.8+ installed on your system
• Basic familiarity with virtual environments
• Access keys for external services used by agents (OpenRouter, Tavily, Firecrawl, ElevenLabs, Alpha Vantage, etc.)

Installation steps:

1. Clone the repository (if not already):

git clone https://github.com/jenasuraj/Ai_agents.git

2. Navigate to the repository or the relevant projects folder that contains the MCP setup.

cd Ai_agents

3. Create a Python virtual environment and activate it:

python -m venv venv
# Windows
venv\Scripts\Activate
# macOS/Linux
source venv/bin/activate

4. Install required Python packages (if a requirements file exists). If not, ensure uvx tooling is installed in your environment:

pip install uvx

5. Install or configure the MCP server package via uvx (as suggested):

uvx jenasauraj-ai_agents

6. Set up necessary API keys and environment variables (see env section below). Then start the MCP server according to your environment (the uvx runner will handle the execution based on the provided package name).

Notes:

• If you prefer a different environment manager or container, you can adapt steps to your setup, but the essential idea is to install the package (jenasuraj-ai_agents) with uvx and provide required API keys via env vars.

Environment variables and configuration:

• OPENROUTER_API_KEY: Required if the Ai_agents rely on OpenRouter for LLM calls.
• LANGGRAPH_API_KEY: If your deployment uses a managed LangGraph service, provide its API key.
• Tool-specific keys (e.g., Tavily, Firecrawl, ElevenLabs, Alpha Vantage, NSE, MoneyControl, GitHub API, Notion API) should be supplied as environment variables or within a secured config file as appropriate for your deployment.

Common issues:

• API keys missing or invalid will cause tool invocations to fail; ensure keys are set and accessible to the MCP runtime.
• Version mismatches between LangGraph/LangChain and MCP runtime can cause tool orchestration failures; pin compatible versions and consult the project docs for supported ranges.
• If running in a restricted environment, ensure outbound network access is allowed for external services (APIs, TTS, scraping endpoints).

Configuration options:

• You can adjust the mcpServers entry to include additional environment variables or override command/args if you fork the server package.
• Monitor logs from the MCP runtime to fine-tune agent routing and tool usage per your use-case.