Tools are the capabilities you can give to an Agent to help it achieve its goal. Just like you would give a person a set of tools to do a job, you can equip an agent with the right functions to interact with the world and perform actions. Tools are what make agents powerful, allowing them to go beyond simple text generation and take meaningful steps to complete tasks.

​
Built-in Tools

Ag.dev comes with a set of pre-built tools that you can enable for any agent with a single click. These tools cover common use cases for information retrieval.
  • Web Search: Allows the agent to perform web searches to find up-to-date information on any topic. This is essential for research tasks.
  • News Search: Gives the agent the ability to search for recent news articles, perfect for tasks that require the latest headlines and developments.
  • Browse Website: Enables the agent to navigate to a specific URL, read its content, and extract information. This is useful for summarizing articles, extracting data from a page, or monitoring a website for changes.
You can select any combination of these tools when creating or editing an agent.

​
MCP (Model Context Protocol) Tools

While built-in tools are powerful, the true potential of agents is unlocked through custom tools. This is where the Model Context Protocol (MCP) system comes in. MCP is an open standard designed to let language models interact seamlessly with external tools and data sources. It acts as a universal bridge, allowing you to connect your own services to the Ag.dev platform securely and reliably. Here’s how it works:
  1. Create an MCP Server: An MCP server is an external service you build that exposes a set of custom tools via the Model Context Protocol. For example, you could create an MCP server that connects to your company’s internal CRM, a proprietary database, or a third-party API.
  2. Attach the Server: In the Ag.dev Console, you can attach your MCP server to an agent by providing its URL.
  3. Enable Tools: Once attached, you can see all the tools your MCP server provides and select which ones the agent is allowed to use.
This extensible system means you can create agents with highly specialized capabilities tailored to your specific workflows. For example, you could build an agent that can:
  • Look up customer information in your CRM.
  • Analyze sales data from a private database.
  • Post messages to a Slack channel.
By building your own MCP servers, you can transform agents from general-purpose assistants into highly specialized, automated team members.