Agent

Understand what an Agent is in Buda and how it relates to Drive, Channels, Sessions, and Skills.

An Agent is the core working unit in Buda. It combines identity, instructions, files, tools, and connected channels into one operational AI worker.

What an Agent includes

Part	Role
Identity	Name, description, and role of the Agent
Instructions	System prompt, behavior rules, and operating constraints
Drive	Private SSD workspace (`/agent`) the Agent can reference
Skills	Extra capabilities the Agent can call when needed
Channels	External communication surfaces bound to the Agent
Sessions	Independent conversation contexts created as users interact

What each Agent means for billing

Each Agent you purchase includes three things:

One human member slot — invite one team member per Agent
One dedicated compute sandbox — isolated CPU + RAM + SSD (/agent directory)
Monthly credits added to the Space pool — not a per-Agent cap; credits accumulate at the Space level

Example: A Pro Space with 5 Agents gets 5 × 20,000 = 100,000 credits/month shared across all Agents.

How an Agent fits into the product

Workspace / Space
    -> Agent
        -> Drive
        -> Skills
        -> Channels
        -> Chat Sessions

One workspace can have multiple Agents. Each Agent can focus on a different job, such as customer support, finance ops, recruiting, or investment research.

When to create a new Agent

Create a separate Agent when you need:

A different knowledge base
A different tone or role
Different channel routing
Different permissions or skills
Cleaner isolation between teams or workflows

Good Agent design

Keep one Agent focused on one job
Give it a clear name and mission
Upload only the files it should actually use
Add channels after the core behavior is already stable
Install skills deliberately instead of turning everything on