What is an AI agent?
An AI agent is a language model that pursues a goal in steps, calling tools, reading their results and deciding what to do next.
How it works
Where a chatbot answers, an agent acts: it can search, read files, write documents or call services, then use what came back to plan its next step, until the task is done or it needs a person.
Anthropic distinguishes workflows, where code decides the steps, from agents, where the model decides them. Either way, an agent is only as safe as the tools and permissions it is given.
An example
Asked to prepare a release note, an agent searches the documents changed this week, reads each one, drafts the note, files it under Releases and links it to the roadmap, then reports what it did.
Common mistakes
- Giving an agent more access than the person it works for.
- Letting it write with no history to compare and restore.
- Trusting instructions found inside documents as if a person had given them.
In Nodary
Two kinds of agents work in Nodary: Studio AI, built in, and the agents you connect over MCP, such as Claude, ChatGPT or Cursor. Both act with the permissions of the person behind them, every write keeps the version before it, and Nodary tells agents that document text is content to work with, not instructions to follow.
What Nodary does not do: agents cannot share links, change access, members, roles, keys or billing, or delete a project: those stay human decisions.
Connect an AI agent · Studio AI and Comets
Try it on your own documents. The Free plan needs no card.
Create a free workspace