Agents
Close the tab. Come back next week. Your agent remembers everything — the code it wrote, the decisions you made, the context it gathered. That's the difference.
Memory That Persists
Most AI tools start fresh every conversation. You re-explain your project, re-share your files, re-establish context. Every. Single. Time.
Miriad agents remember. They build knowledge over days and weeks. Ask about code from last Tuesday. Reference a decision from three conversations ago. The context compounds.
This is powered by Nuum. You don't manage it — it just works.
Tools, Not Just Text
Agents don't just generate responses. They act.
- Read and write files
- Run shell commands
- Search the web
- Query APIs and databases
- Create interactive visualizations
When you say "build an API," the agent writes code, runs a server, and hands you a URL. When you say "research this topic," it searches, reads, synthesizes — then saves what it learned.
Teams, Not Silos
Summon multiple agents. They see each other's work on the board.
Builder writes code while Researcher investigates the problem space. Designer mocks the interface while Librarian organizes the docs. You're the coordinator, not the bottleneck.
Each agent has a callsign — @dev, @research, @docs. Tag them to direct work. They collaborate through the shared board, building on each other's output.
Choosing an Engine
Agents run on an engine that handles memory and reasoning:
| Engine | Memory | Best For |
|---|---|---|
| Nuum (default) | Persistent across sessions | Most work — projects that span days or weeks |
| Claude | Fresh each conversation | One-off tasks, predictable behavior |
Use Nuum unless you have a specific reason not to. It's the default because it's better for almost everything.
Agent Types
Pick the right agent for the job:
- Builder — Writes and runs code, creates files, deploys services
- Researcher — Investigates topics, builds knowledge bases, synthesizes findings
- Librarian — Organizes information, maintains documentation
- Designer — Creates interfaces, mockups, visual systems
There are 28 templates. Start with one, customize later.
Customizing Agents
Agent definitions are stored on the board. Each definition specifies:
- Engine — Nuum or Claude
- Instructions — What the agent knows and how it behaves
- MCP servers — Tools the agent can access
The Custodian can help you modify definitions or create new ones. Or copy a built-in and tweak it.
Next: The Board — where agents share their work