Memory

1. What Is Memory?

Memory is the agent’s long‑term understanding of your world. It’s how agents stay consistent, continue threads, and build on previous work.

In Playnex, memory is not stored inside the model. Instead, it lives in capsules — structured containers that agents can read, write, and reference.

This makes memory:

• Transparent — you can see exactly what agents remember
• Editable — you can refine or remove memory entries
• Portable — memory syncs across your devices
• Private — stored locally unless you choose otherwise

---

2. How Agents Use Memory

Agents read from memory when they need context, and write to memory when they produce something worth keeping.

For example, an agent might:

• Save a research summary into a capsule
• Store a draft it generated
• Record a plan or outline
• Capture insights from a conversation

Over time, these capsules form a growing archive — a map of your thinking and your agent’s contributions.

---

3. Viewing Memory

To browse all memory capsules:

playnex capsules

To view a specific memory entry:

playnex capsule view <id>

You can also view and edit memory in the Playnex dashboard.

---

4. Writing to Memory

Agents can write to memory automatically, or you can instruct them explicitly:

“Save this as a memory capsule.”

Or from the CLI:

playnex capsule create

Memory entries are just capsules — but capsules that agents treat as long‑term context.

---

5. Editing and Deleting Memory

Because memory is transparent, you can refine it at any time.

To delete a memory capsule:

playnex capsule delete <id>

Agents will immediately stop referencing it.

Memory is not magic. It’s a clear, human‑readable layer that grows with you — a shared space between your intelligence and your agents’.