|
9 | 9 | "\n", |
10 | 10 | "Most agents start every conversation from scratch. A customer tells your shopping assistant their size, their budget, and which materials they avoid, and the next time they return, the agent has forgotten everything. They have to repeat themselves, and the experience feels generic rather than personal.\n", |
11 | 11 | "\n", |
12 | | - "We just introduced Memory in Claude Managed Agents to solve this. Think of it as a shared notebook your agent gets for each customer: Claude jots relevant things down during a session, and those notes are still there the next time the same customer comes back. Setup is a simple API call away.\n", |
| 12 | + "We just introduced [Memory](https://platform.claude.com/docs/en/managed-agents/memory) in Claude Managed Agents to solve this. Think of it as a shared notebook your agent gets for each customer: Claude jots relevant things down during a session, and those notes are still there the next time the same customer comes back. Setup is a simple API call away.\n", |
13 | 13 | "\n", |
14 | 14 | "In this guide you will build an example shopping assistant for a retail brand. The agent will learn a customer's preferences during their first visit, save them to a user-specific memory store, and recall them automatically on the next visit without being told again.\n", |
15 | 15 | "\n", |
|
618 | 618 | "\n", |
619 | 619 | "From here you can map your own user IDs to memory store IDs, seed stores from your existing customer data, and layer shared read-only stores on top for brand-wide knowledge.\n", |
620 | 620 | "\n", |
621 | | - "### Other notebooks in this series\n", |
622 | | - "\n", |
623 | | - "- [`CMA_iterate_fix_failing_tests.ipynb`](CMA_iterate_fix_failing_tests.ipynb) — the entry-point notebook. Introduces agents, environments, sessions, file mounts, and the streaming event loop through a do-observe-fix loop on a failing test suite.\n", |
624 | | - "- [`CMA_operate_in_production.ipynb`](CMA_operate_in_production.ipynb) — production setup story: vault-backed MCP credentials, the `session.status_idled` webhook for HITL without long-lived connections, and the resource lifecycle CRUD verbs.\n", |
625 | | - "\n", |
626 | 621 | "### Learn more\n", |
627 | 622 | "\n", |
628 | 623 | "- [Claude Managed Agents overview](https://docs.anthropic.com/en/docs/managed-agents/overview)\n", |
629 | | - "- [Memory stores API reference](https://docs.anthropic.com/en/docs/managed-agents/memory)\n", |
| 624 | + "- [Memory stores API reference](https://platform.claude.com/docs/en/managed-agents/memory)\n", |
630 | 625 | "- [Session resources](https://docs.anthropic.com/en/docs/managed-agents/memory#attach-a-memory-store-to-a-session)" |
631 | 626 | ] |
632 | 627 | } |
|
0 commit comments