Skip to content

Module 5: Agent Memory

Give your agent memory. Session context tracks the current conversation. Long term preferences persist across conversations.

How it works

Memory type What it stores Lifetime
Session Every message in the current conversation Expires after the session ends
Long term User preferences and facts Persists across conversations

Cloud setup

1. Create an Agent Memory store

  1. In the Redis Cloud console, find Agent Memory in the left sidebar.

  2. Click Quick Create.

    Agent Memory Quick Create

2. Copy your API key

Save this — you won't see it again.

Agent Memory service key

3. Get your Endpoint and Store ID

After dismissing the key dialog, find both on the store details page:

Agent Memory credentials

4. Configure environment

Add to your .env:

MEMORY_API_BASE_URL=<endpoint from step 3>
MEMORY_STORE_ID=<store-id from step 3>
MEMORY_API_KEY=<API key from step 2>

5. Seed long term memories

make seed-memories

.\workshop.ps1 seed-memories

This seeds two preferences for the demo customer: "Prefers paperless statements and online banking" and "Interested in fixed deposit products for savings growth."

Exercise

Open exercises/banking/agent_memory.py

long_term_search_payload(text, owner_id, session_id, limit)

This tells Agent Memory how to search for stored user preferences. When a conversation starts, the agent calls this to recall what it knows about the user.

def long_term_search_payload(self, *, text, owner_id, session_id, limit):
    return {
        "text": text,                                        # what to search for
        "similarityThreshold": 0.2,                          # how close a match must be
        "filterOp": "all",                                   # all filters must match
        "limit": limit or 5,                                 # max memories to return
        "filter": {
            "ownerId": {"eq": sanitize_owner_id(owner_id)},  # scope to this user
            "namespace": {"eq": "banking-demo"},          # scope to this app
        },
    }

Fill in the method body with the dict above.

  • filter scopes results so each user only sees their own memories
  • sanitize_owner_id is already imported at the top of the file

Try it

Restart with make dev, then open localhost:3040.

  1. Ask: "What are my banking preferences and savings interests?" — the agent recalls "paperless statements" and "interested in fixed deposits"
  2. Ask a follow up: "What are the current rates for those?" — the agent remembers what you just discussed

Verify

  • Agent recalls seeded preferences ("paperless statements and online banking" and "fixed deposit products for savings growth")
  • Multi turn conversations maintain context
  • Activity panel shows session and long term memory lookups