[New Example] Long Term Memory for Gemini #987

lucifertrj · 2025-10-08T21:37:08Z

Save the preference in the Mem0.AI for the memory layer
Based on the given user query, search and retrieve similar preferences from the memory layer.
This example is use case driven that is useful for the recommendation systems and travel or food place planner.
It uses Gemini as LLM, Qdrant as Vector Store and FastEmbed as Embeddings (open source)

gemini-code-assist · 2025-10-08T21:37:14Z

Note

Gemini is unable to generate a summary for this pull request due to the file types involved not being currently supported.

review-notebook-app · 2025-10-08T21:37:14Z

Check out this pull request on

See visual diffs & provide feedback on Jupyter Notebooks.

Powered by ReviewNB

Giom-V · 2025-10-08T22:57:50Z

/gemini review

gemini-code-assist · 2025-10-08T22:57:55Z

Note

Gemini is unable to generate a review for this pull request due to the file types involved not being currently supported.

lucifertrj

do you need it to be in ipynb or should I replace it with py for Gemini code assistant to perform the review.

chunduriv · 2025-10-09T21:32:21Z

Hi @lucifertrj, thanks for the contribution!

Please keep the example as a Interactive Python Notebook (.ipynb). Notebooks are the standard format for examples in this cookbook.

It looks like the notebook file isn't rendering correctly on GitHub due to a metadata issue.

Could you please try clearing all the cell outputs in your notebook and then pushing the updated file to this PR? That should resolve the rendering issue.

If you're still having trouble, feel free to share a Colab Gist so I can review your work in the meantime.

lucifertrj · 2025-10-12T02:18:21Z

sure will modify

lucifertrj · 2025-10-12T02:30:02Z

Hey @chunduriv I have removed the cell outputs.

you can view the notebook here: https://github.com/lucifertrj/cookbook/blob/longtermmemory/examples/Long_Term_Memory.ipynb

lucifertrj · 2025-10-12T02:30:48Z

For the GitHub gist, you can refer to this: https://colab.research.google.com/gist/lucifertrj/91c7fc265a6d40733c7a99bcf9aaa31d/long_term_memory.ipynb

lucifertrj · 2025-10-23T15:35:19Z

@Giom-V and @chunduriv any updates? I have formatted the notebook -nbfmt.

Giom-V · 2025-10-31T16:26:16Z

examples/Long_Term_Memory.ipynb

@@ -0,0 +1,465 @@
+{


It's better to use %pip in notebooks

Reply via ReviewNB

Giom-V · 2025-10-31T16:26:16Z

examples/Long_Term_Memory.ipynb

@@ -0,0 +1,465 @@
+{


Can you use GEMINI_API_KEY instead? We're going to migrate to this new naming soon.

Reply via ReviewNB

Giom-V · 2025-10-31T16:26:16Z

examples/Long_Term_Memory.ipynb

@@ -0,0 +1,465 @@
+{


Why not the gemini embedding model

Reply via ReviewNB

I will switch to gemini embedding

Giom-V · 2025-10-31T16:26:16Z

examples/Long_Term_Memory.ipynb

@@ -0,0 +1,465 @@
+{


It's unclear to me from the code samples what was saved to memory (all the conversation?). Can you add more explanations?

Reply via ReviewNB

Giom-V · 2025-10-31T16:26:16Z

examples/Long_Term_Memory.ipynb

@@ -0,0 +1,465 @@
+{


I think it would be great to explain what's happening here. Is it searching in the memory?

Reply via ReviewNB

sure, I will add the information with one flowchart

Giom-V · 2025-10-31T16:26:16Z

examples/Long_Term_Memory.ipynb

@@ -0,0 +1,465 @@
+{


Can you save the outputs in the notebook so one can see what it does without running it

Reply via ReviewNB

actually It was present, previously it was mentioned to clear the output cells, so I had removed.

Giom-V · 2025-10-31T16:26:16Z

examples/Long_Term_Memory.ipynb

@@ -0,0 +1,465 @@
+{


Can you indent the text like this?

SYSTEM_PROMPT = """ You are an expert executive assistant who thinks carefully before responding, adapting to the poliet communication style based on the previous user's established PREFERENCES and the complexity of their query. Maintain a polished, professional tone that is warm yet efficient—concise for simple questions, moderate for complex topics, and comprehensive for open-ended discussions. Act as a trusted advisor who doesn't just answer questions but adds value through insights, anticipates needs, and prioritizes what matters most while respecting the user's time with clear, actionable responses. """

Reply via ReviewNB

Giom-V · 2025-10-31T16:29:44Z

@lucifertrj I left you some comments. Basically add more explanations about what's happening.
Can you also add a "what's next?" section at the end?
I'll just try to run the gemini review, but take it with a grain of salt, it can be a bit picky.

Giom-V · 2025-10-31T16:29:51Z

/gemini review

gemini-code-assist

Code Review

This pull request introduces a new example notebook demonstrating long-term memory for Gemini using Mem0.AI, Qdrant, and FastEmbed. The example is a valuable addition, showcasing a practical use case for recommendation systems. However, the notebook has several areas that need improvement to align with the repository's style guide. Key issues include incorrect markdown formatting (title, headings, Colab badge URL), use of !pip instead of %pip, hardcoded model names instead of using Colab forms, and formatting issues like line length and inconsistent code style. Additionally, the notebook appears to have not been fully run and formatted before submission, as indicated by populated execution_count fields and missing cell outputs. Finally, this new notebook needs to be added to the examples/README.md file.

gemini-code-assist · 2025-10-31T16:32:14Z

examples/Long_Term_Memory.ipynb

+      "source": [
+        "## Long Memory Layer using - Mem0, Gemini and Qdrant\n",
+        "\n",
+        "<a target=\"_blank\" href=\"https://colab.research.google.com/drive/13PYVsCenlKOI2iUnPE0A_LcfGDElOWTR?usp=sharing\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" height=30/></a>"


The URL for the 'Open in Colab' badge is incorrect. It should point to the notebook's location within the GitHub repository, not a Google Drive link, to allow users to open it directly.¹

Suggested change

"<a target=\"_blank\" href=\"https://colab.research.google.com/drive/13PYVsCenlKOI2iUnPE0A_LcfGDElOWTR?usp=sharing\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" height=30/></a>"

"<a target=\"_blank\" href=\"https://colab.research.google.com/github/google-gemini/cookbook/blob/main/examples/Long_Term_Memory.ipynb\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" height=30/></a>"

Style Guide References

Footnotes

The style guide provides a specific format for the 'Open in Colab' badge URL, which should point to the notebook on GitHub. ↩

gemini-code-assist · 2025-10-31T16:32:14Z

examples/Long_Term_Memory.ipynb

+        "id": "4L89T6n6qibs"
+      },
+      "source": [
+        "## Long Memory Layer using - Mem0, Gemini and Qdrant\n",


The main title of the notebook should be a level 1 heading (H1) and use sentence case, as per the style guide.¹²

Suggested change

"## Long Memory Layer using - Mem0, Gemini and Qdrant\n",

"# Long memory layer using Mem0, Gemini and Qdrant\n",

Style Guide References

Footnotes

The style guide requires using one H1 header (# in Markdown) for the title. ↩

The style guide specifies using sentence case for titles and headings. ↩

gemini-code-assist · 2025-10-31T16:32:14Z

examples/Long_Term_Memory.ipynb

+    },
+    {
+      "cell_type": "code",
+      "execution_count": 1,


The execution_count for this cell is 1, and for many others it is not null. This indicates that the notebook formatting script has not been run. Please run the formatter on the notebook before submitting to clear execution counts and ensure consistent formatting.¹

Style Guide References

Footnotes

If the execution_count has changed to something other than null, it usually indicates that the formatting script has not been run. A warning should be raised. ↩

gemini-code-assist · 2025-10-31T16:32:14Z

examples/Long_Term_Memory.ipynb

+        "!pip install agno mem0ai google-genai\n",
+        "!pip install langchain langchain-community fastembed"


The style guide requires using %pip instead of !pip for installations. It's also recommended to combine installations into a single command and use the -U -q flags for google-genai as specified in the guide.¹²

Suggested change

"!pip install agno mem0ai google-genai\n",

"!pip install langchain langchain-community fastembed"

"%pip install -U -q agno mem0ai 'google-genai>=1.0.0' langchain langchain-community fastembed"

Style Guide References

Footnotes

The style guide specifies using %pip instead of !pip. ↩

The style guide specifies the correct way to install the google-genai SDK, including version and quiet flags. ↩

gemini-code-assist · 2025-10-31T16:32:14Z

examples/Long_Term_Memory.ipynb

+        "id": "3FSrgmEPuC4R"
+      },
+      "source": [
+        "### Define the Memory Configuration\n",


Headings should use sentence case as per the style guide.¹

Suggested change

"### Define the Memory Configuration\n",

"### Define the memory configuration\n",

Style Guide References

Footnotes

The style guide requires using sentence case in titles/headings. ↩

gemini-code-assist · 2025-10-31T16:32:15Z

examples/Long_Term_Memory.ipynb

+      },
+      "outputs": [],
+      "source": [
+        "result1"


This cell is expected to display the content of result1, but its output is missing. To improve readability, notebooks should be saved with their outputs so users can see the results without running the code themselves.¹ Please run the notebook to completion and save it with the outputs included. This also applies to other cells where output is expected but missing (e.g., lines 298, 351, 449).

Style Guide References

Footnotes

Ideally we want the ouputs to be saved in the notebooks so that one can see what the code does without runnning it. ↩

examples/Long_Term_Memory.ipynb

gemini-code-assist · 2025-10-31T16:32:15Z

examples/Long_Term_Memory.ipynb

+      "source": [
+        "SYSTEM_PROMPT = \"\"\"\n",
+        "You are an expert executive assistant who thinks carefully before responding,\n",
+        "adapting to the poliet communication style based on the previous user's established PREFERENCES and the complexity of their query.\n",


There is a typo in the word "poliet". It should be "polite".

adapting to the polite communication style based on the previous user's established PREFERENCES and the complexity of their query.\n,

gemini-code-assist · 2025-10-31T16:32:15Z

examples/Long_Term_Memory.ipynb

+        "    USER_PROMPT = f\"\"\"\n",
+        "      <question>\n",
+        "      QUESTION: {query}\n",
+        "      </question>\n",
+        "\n",
+        "      <PREFERENCE>\n",
+        "      Preference: {context}\n",
+        "      </PREFERENCE>\n",
+        "    \"\"\"\n",


The indentation in this multi-line f-string is inconsistent. Please format it according to the style guide for long strings to improve readability.¹

USER_PROMPT = f""" <question> QUESTION: {query} </question> <PREFERENCE> Preference: {context} </PREFERENCE> """

Style Guide References

Footnotes

Long text variables should use triple double quotes and proper indentation for better readability. ↩

gemini-code-assist · 2025-10-31T16:32:15Z

examples/Long_Term_Memory.ipynb

+        "\n",
+        "    # Config the system prompt and make sure to define the input variables inside the USER PROMPT\n",
+        "    response = llm_client.models.generate_content(\n",
+        "        model=\"gemini-2.5-pro\",\n",


The model ID is hardcoded. For better maintainability and ease of use, the style guide recommends using a Colab form selector to choose the model.¹ Please define the model in a separate cell using a parameter, for example:

MODEL_ID = "gemini-2.5-pro" # @param ["gemini-2.5-flash", "gemini-2.5-pro"]

Then, use this MODEL_ID variable here.

Style Guide References

Footnotes

When selecting a model, use a colab selector for easier maintainability. ↩

lucifertrj · 2025-10-31T18:56:53Z

@lucifertrj I left you some comments. Basically add more explanations about what's happening. Can you also add a "what's next?" section at the end? I'll just try to run the gemini review, but take it with a grain of salt, it can be a bit picky.

sure. will make the modifications

lucifertrj · 2025-11-10T20:13:16Z

@Giom-V I have resolved the comments:

replace the !pip install to %pip install
replace GOOGLE_API_KEY to GEMINI_API_KEY
Switched the embedding model to Google embedding model (previously it was FastEmbed)
Added more information details for memory add operation and search operation
Removed the dependency of langchain, earlier we used it for fastembed, now its not needed.

long term memory example for gemini

82e2406

github-actions bot added status:awaiting review PR awaiting review from a maintainer component:examples Issues/PR referencing examples folder labels Oct 8, 2025

chunduriv self-assigned this Oct 8, 2025

lucifertrj commented Oct 9, 2025

View reviewed changes

chunduriv added status:awaiting response Awaiting a response from the author and removed status:awaiting review PR awaiting review from a maintainer labels Oct 9, 2025

clear cell outputs

82ecd79

chunduriv requested a review from Giom-V October 14, 2025 00:21

format notebook and fix lint

c3a592e

Giom-V reviewed Oct 31, 2025

View reviewed changes

gemini-code-assist bot reviewed Oct 31, 2025

View reviewed changes

make changes as per the comments added

3e1ac56

lucifertrj force-pushed the longtermmemory branch from d57677e to 3e1ac56 Compare November 10, 2025 20:07

	"<a target=\"_blank\" href=\"https://colab.research.google.com/drive/13PYVsCenlKOI2iUnPE0A_LcfGDElOWTR?usp=sharing\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" height=30/></a>"
	"<a target=\"_blank\" href=\"https://colab.research.google.com/github/google-gemini/cookbook/blob/main/examples/Long_Term_Memory.ipynb\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" height=30/></a>"

	"## Long Memory Layer using - Mem0, Gemini and Qdrant\n",
	"# Long memory layer using Mem0, Gemini and Qdrant\n",

		"!pip install agno mem0ai google-genai\n",
		"!pip install langchain langchain-community fastembed"

	"!pip install agno mem0ai google-genai\n",
	"!pip install langchain langchain-community fastembed"
	"%pip install -U -q agno mem0ai 'google-genai>=1.0.0' langchain langchain-community fastembed"

	"### Define the Memory Configuration\n",
	"### Define the memory configuration\n",

[New Example] Long Term Memory for Gemini #987

Are you sure you want to change the base?

[New Example] Long Term Memory for Gemini #987

Uh oh!

Conversation

lucifertrj commented Oct 8, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

gemini-code-assist bot commented Oct 8, 2025

Uh oh!

review-notebook-app bot commented Oct 8, 2025

Uh oh!

Giom-V commented Oct 8, 2025

Uh oh!

gemini-code-assist bot commented Oct 8, 2025

Uh oh!

lucifertrj left a comment

Choose a reason for hiding this comment

Uh oh!

chunduriv commented Oct 9, 2025

Uh oh!

lucifertrj commented Oct 12, 2025

Uh oh!

lucifertrj commented Oct 12, 2025

Uh oh!

lucifertrj commented Oct 12, 2025

Uh oh!

lucifertrj commented Oct 23, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Giom-V Oct 31, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Giom-V Oct 31, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Giom-V Oct 31, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

lucifertrj Oct 31, 2025

Choose a reason for hiding this comment

Uh oh!

Giom-V Oct 31, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Giom-V Oct 31, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

lucifertrj Oct 31, 2025

Choose a reason for hiding this comment

Uh oh!

Giom-V Oct 31, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

lucifertrj Oct 31, 2025

Choose a reason for hiding this comment

Uh oh!

Giom-V Oct 31, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Giom-V commented Oct 31, 2025

Uh oh!

Giom-V commented Oct 31, 2025

Uh oh!

gemini-code-assist bot left a comment

Choose a reason for hiding this comment

Code Review

Uh oh!

gemini-code-assist bot Oct 31, 2025

Choose a reason for hiding this comment

Style Guide References

Footnotes

Uh oh!

lucifertrj commented Oct 8, 2025 •

edited

Loading

lucifertrj commented Oct 23, 2025 •

edited

Loading

Giom-V Oct 31, 2025 •

edited

Loading

Giom-V Oct 31, 2025 •

edited

Loading

Giom-V Oct 31, 2025 •

edited

Loading

Giom-V Oct 31, 2025 •

edited

Loading

Giom-V Oct 31, 2025 •

edited

Loading

Giom-V Oct 31, 2025 •

edited

Loading

Giom-V Oct 31, 2025 •

edited

Loading