> For the complete documentation index, see [llms.txt](https://oodlz-ai-studio.gitbook.io/docs/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://oodlz-ai-studio.gitbook.io/docs/inbound-agent/every-channel/one-agent-one-memory.md).

# One Agent, One Memory

One session key — \`{platform}\_{senderId}\` — shared across every channel, so a user who switches from Instagram DM to the website widget picks up the same thread.

> **A user who starts on Instagram and switches to the website doesn't restart the conversation.**

Stateless chatbots forget the moment the conversation ends. Ask one "what did I just ask about?" and it has no idea. The Oodlz inbound agent maintains a 10-message rolling memory window per user with a 24-hour TTL — and the session-key scheme makes that memory portable across every channel.

## How It Works

Each session is keyed `{platform}_{senderId}`. The same scheme is used by the website widget, the social listening pipeline, and the voice client. When a user opens the website widget after sending an Instagram DM, the agent looks up the same key, finds the 10-message context, and continues the thread. The agent can recall a brand mentioned in a previous turn or pick up where the user left off without any extra session bookkeeping in the calling client.

Memory expires 24 hours after the last turn — long enough to span a normal customer interaction across channels, short enough that abandoned sessions don't accumulate.

## What You Get

* One conversation thread per user, regardless of which channel they switch to.
* 10-message rolling context — the agent has the recent history, not just the current message.
* No restated context required — a user mentioning brands from a previous turn is understood without restating which ones.
* 24-hour memory window — automatic cleanup, no manual session management.