Not a feature dump — a way in to understanding agents
Three tracks, one goal: help you understand AI agents and manage them with confidence. In-depth articles are landing here next, each citing primary sources.
Understand AI agents
In plain words and trusted sources: how an agent differs from a chat model, and its core capabilities.
- What is an AI agent?
A chatbot answers; an agent acts. Plain-language definition of an AI agent and how it differs from a chat model.
- The six core capabilities of an agent
Perception, reasoning & planning, tool use, memory, autonomous multi-step execution, and multi-agent collaboration — explained simply.
- Large model vs. agent: what actually changes
The same model, with or without an agent around it, behaves very differently. Here is the distinction that matters.
Berth features in depth
A walk through Overview, Sessions, Configuration and Usage — and the asset model behind them.
- The asset model: what Berth actually shows you
Berth turns the plain-text files behind your agents into structured, connected objects it calls assets. Here is the model.
- Overview & Sessions: see activity and history
The dashboard at a glance, and how to walk back through past sessions with the assets and tools each one used.
- Configuration · Instructions: memories, skills, subagents
The instruction assets that guide your agent — and how Berth shows their scope, imports, and where each one comes from.
- Configuration · Capabilities: MCP, hooks, permissions
The capability assets that give your agent power and set its boundaries — MCP servers, lifecycle hooks, and permissions.
- Usage, health checks & privacy
Cost and token trends, automated diagnostics, and the read-only / local-first guarantees behind it all.
Hands-on guides
Diagnose why a hook isn't firing, make sense of your cost, set a config baseline for your team.
- Why isn’t my hook firing?
A short checklist to diagnose a hook that never runs — using what Berth shows you.
- Make sense of your cost
Read Berth’s Usage screen to find what’s expensive and why — by model, project, and day.
- Set a config baseline for your team
Use scope and imports to give a team a shared, predictable agent setup — and verify it with health checks.