MCP-Server-TESS

MCP (Model Context Protocol) server for integration with the TESS API.

On

This project implements a server that follows the MCP protocol to interact with the TESS API. The server exposes tools that allow:

Listing and Managing Agents
Running agents with custom messages
Managing files and their associations with agents
And much more

Requirements

Node.js 18+
A TESS platform API key

Installation

Via Smithery.ai (Recommended)

You can use this server directly in Smithery.ai:

Visit https://smithery.ai/server/@diegofornalha/mcp-server-tess
Click "Install"
Configure your TESS API key when prompted
Ready! The server is available for use with your favorite LLM

Local Installation

Clone the repository and install the dependencies:

git clone https://github.com/seu-usuario/mcp-server-tess.git
cd mcp-server-tess
npm install

Settings

Create a .env file based on .env.example
Add your TESS API key:

TESS_API_KEY=sua_chave_api_aqui
PORT=3000

Compilation

npm run build

Execution

To start the server in production mode:

npm start

For development with automatic reloading:

npm run dev

Available tools

The server exposes the following tools via HTTP API:

listar_agentes_tess - List all available agents
obter_agente_tess - Gets details of a specific agent
executar_agente_tess - Runs an agent with custom messages
listar_arquivos_agente_tess - List files associated with an agent
vincular_arquivo_agente_tess - Links a file to an agent
remover_arquivo_agente_tess - Removes the link between a file and an agent
listar_arquivos_tess - List all available files
obter_arquivo_tess - Gets details of a specific file
enviar_arquivo_tess - Sends a new file to the TESS platform
excluir_arquivo_tess - Deletes a file from the TESS platform

Usage with Docker

To run the server using Docker:

# Construir a imagem
docker build -t mcp-server-tess .

# Executar o container
docker run -p 3000:3000 -e TESS_API_KEY=sua_chave_api_aqui mcp-server-tess

API Endpoints

GET /health - Server health check endpoint
GET /capabilities - Lists all available tools with their descriptions and parameters
POST /tools/:toolName - Runs a specific tool (replace :toolName with the name of the tool)

Usage example

With Smithery.ai

Once installed on Smithery.ai, you can use the server with any MCP-compatible LLM:

Select the @diegofornalha/mcp-server-tess server in your LLM settings
TESS API tools will automatically be available for your model

Via HTTP API

To run an agent:

curl -X POST http://localhost:3000/tools/executar_agente_tess \
  -H "Content-Type: application/json" \
  -d '{
    "agent_id": "seu_agent_id",
    "model": "tess-ai-light",
    "messages": [
      {"role": "user", "content": "Olá, como você está?"}
    ]
  }'

License

MIT

Integrations