Agents on Apps: langgraph long term agent memory example #86

jennsun · 2026-01-16T23:28:09Z

Added long term memory/retrieval features to our agent langgraph as a stateful agent example

Note: frontend e2e-chatbot-app interface does not support passing in custom input/thread id, only invocation calls on the endpoint. Have filed a ticket to support passing in chatcontext via frontend interface so we can use user id for namespacing

example app: https://eng-ml-agent-platform.staging.cloud.databricks.com/apps/j-longtermagent?o=2850744067564480

Example local testing:

curl -X POST http://localhost:8000/invocations \
  -H "Content-Type: application/json" \
  -d '{
    "input": [{"role": "user", "content": "Remember that my favorite color is blue"}],
    "context": {"user_id": "[email protected]"}
  }'

output from above:

{"object":"response","output":[{"type":"function_call","id":"lc_run--019b059a-09ec-79c3-bcf2-640edb80d37a","call_id":"toolu_bdrk_013tUwgLNctjrTSSSUBWShXQ","name":"save_user_memory","arguments":"{\"memory_key\": \"favorite_color\", \"memory_data_json\": \"{\\\"favorite_color\\\": \\\"blue\\\"}\"}"},{"type":"function_call_output","call_id":"toolu_bdrk_013tUwgLNctjrTSSSUBWShXQ","output":"Successfully saved memory with key 'favorite_color' for user."},{"type":"message","id":"lc_run--019b059a-20c5-76a1-be57-308598158d43","content":[{"text":"Got it! I've remembered that your favorite color is blue. 💙","type":"output_text","annotations":[]}],"role":"assistant"}]}

Follow-up:

curl -X POST http://localhost:8000/invocations \
  -H "Content-Type: application/json" \
  -d '{
    "input": [{"role": "user", "content": "What is my favorite color?"}],
    "context": {"user_id": "[email protected]"}
  }'

Output:

{"object":"response","output":[{"type":"function_call","id":"lc_run--019b059e-a723-76f3-98cb-46737a1ae5f4","call_id":"toolu_bdrk_01QXVHEFgp3VedAuc63HW1b6","name":"get_user_memory","arguments":"{\"query\": \"favorite color\"}"},{"type":"function_call_output","call_id":"toolu_bdrk_01QXVHEFgp3VedAuc63HW1b6","output":"User memories:\n- {\"favorite_color\": \"blue\"}"},{"type":"message","id":"lc_run--019b059e-cbd9-7252-86ff-2ae05615c341","content":[{"text":"Your favorite color is blue!","type":"output_text","annotations":[]}],"role":"assistant"}]}

seeing store tables are ready in app logs:

ensure user preferences are stored in lakebase:

Currently, cannot use long-term memory via frontend because user id is not passed in to context. Also made sure you cannot access another user's memories via frontend

App postman request:

bbqiu

this PR looks awesome! left a few small nits, thank you for working on this!

bbqiu · 2026-01-17T22:46:46Z

agent-langgraph-long-term-memory/scripts/quickstart.sh

+# Check if LAKEBASE_INSTANCE_NAME already exists in app.yaml
+if grep -q "name: LAKEBASE_INSTANCE_NAME" app.yaml; then
+    echo "LAKEBASE_INSTANCE_NAME already exists in app.yaml, skipping Lakebase setup..."
+    # Extract the existing value from app.yaml for display in summary


do we need this? since app.yaml has valueFrom database instead of a real string value

can also make this an input param to match what sid also is working on making the quickstart command require less user input (so AI can run it fully): https://github.com/databricks/app-templates/pull/82/files#diff-0634e5169824dfbd324f2962bb69f68e1dd4445e4ee953cbe74b04e1ca9b8329

bbqiu · 2026-01-17T22:47:46Z

agent-langgraph-long-term-memory/scripts/quickstart.sh

+    fi
+    echo "✓ Lakebase instance name saved to .env.local"
+
+    echo "Adding LAKEBASE_INSTANCE_NAME to app.yaml..."


do we need to replace the app.yaml value here?

bbqiu · 2026-01-17T22:48:10Z

agent-langgraph-long-term-memory/scripts/quickstart.sh

+echo "✓ Configuration files created (.env.local)"
+echo "✓ MLflow experiment created: $EXPERIMENT_NAME"
+echo "✓ Experiment ID: $EXPERIMENT_ID"
+echo "✓ Configuration updated in .env.local"


nit: can we add a line here for the lakebase instance name

bbqiu · 2026-01-17T22:49:05Z

agent-langgraph-long-term-memory/README.md

+
+1. **Set up authentication to Databricks resources**
+
+   For this example, you need to add an MLflow Experiment and Lakebase instance as a resource to your app. Grant the App's Service Principal (SP) permission to edit the experiment by clicking `edit` on your app home page. See the [Databricks Apps MLflow experiment documentation](https://docs.databricks.com/aws/en/dev-tools/databricks-apps/mlflow) for more information.


nit: can we add link to database docs

bbqiu · 2026-01-17T22:49:43Z

agent-langgraph-long-term-memory/README.md

+   ```
+
+3. **Grant Lakebase permissions to your App's Service Principal**
+


nit: can we explain what's going on here? where connect + create doesn't grant perms for previously created schemas etc.

bbqiu · 2026-01-17T22:51:20Z

agent-langgraph-long-term-memory/agent_server/agent.py

+############################################
+# Memory tools
+############################################
+def memory_tools():


nit: can we name it init_memory_tools to make it clear that this is to create them

agent-langgraph-long-term-memory/agent_server/agent.py

agent-langgraph-long-term-memory/README.md

jennsun added 2 commits January 16, 2026 14:26

copy agent-langgraph into agent-langgraph-long-term-memory

b49e59a

stateful long term updates

af24bd0

bbqiu mentioned this pull request Jan 17, 2026

Agent memory langgraph long-term-memory using semantic search bbqiu/agent-on-app-proto#42

Closed

bbqiu self-requested a review January 17, 2026 22:27

bbqiu approved these changes Jan 17, 2026

View reviewed changes

bbqiu reviewed Jan 17, 2026

View reviewed changes

agent-langgraph-long-term-memory/agent_server/agent.py Show resolved Hide resolved

bbqiu reviewed Jan 17, 2026

View reviewed changes

agent-langgraph-long-term-memory/README.md Outdated Show resolved Hide resolved

jennsun added 2 commits January 19, 2026 16:43

pr review

c561e44

update readme

b36e649

bbqiu merged commit 3e28abf into databricks:main Jan 21, 2026

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Agents on Apps: langgraph long term agent memory example #86

Agents on Apps: langgraph long term agent memory example #86

Uh oh!

jennsun commented Jan 16, 2026 •

edited

Loading

Uh oh!

bbqiu left a comment

Uh oh!

bbqiu Jan 17, 2026

Uh oh!

bbqiu Jan 17, 2026

Uh oh!

bbqiu Jan 17, 2026

Uh oh!

bbqiu Jan 17, 2026

Uh oh!

bbqiu Jan 17, 2026

Uh oh!

bbqiu Jan 17, 2026

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants


		1. Set up authentication to Databricks resources

		For this example, you need to add an MLflow Experiment and Lakebase instance as a resource to your app. Grant the App's Service Principal (SP) permission to edit the experiment by clicking `edit` on your app home page. See the [Databricks Apps MLflow experiment documentation](https://docs.databricks.com/aws/en/dev-tools/databricks-apps/mlflow) for more information.

		```

		3. Grant Lakebase permissions to your App's Service Principal

Agents on Apps: langgraph long term agent memory example #86

Agents on Apps: langgraph long term agent memory example #86

Uh oh!

Conversation

jennsun commented Jan 16, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

bbqiu left a comment

Choose a reason for hiding this comment

Uh oh!

bbqiu Jan 17, 2026

Choose a reason for hiding this comment

Uh oh!

bbqiu Jan 17, 2026

Choose a reason for hiding this comment

Uh oh!

bbqiu Jan 17, 2026

Choose a reason for hiding this comment

Uh oh!

bbqiu Jan 17, 2026

Choose a reason for hiding this comment

Uh oh!

bbqiu Jan 17, 2026

Choose a reason for hiding this comment

Uh oh!

bbqiu Jan 17, 2026

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

jennsun commented Jan 16, 2026 •

edited

Loading