add goodmem save/fetch tools and chat plugin

[forrestbao]

PR: Goodmem tools, plugin, and memory service for ADK

This PR adds Goodmem.ai integrations to ADK: two tools, one plugin for automatic chat memory in conversational agents, and one memory service.

---

Files added / changed (ASCII tree)

adk-python-community/
├── .gitignore                                    (M  – add .adk/ ignore)
├── pyproject.toml                                (M  – add requests, plugins/tools)
├── src/google/adk_community/
│   ├── __init__.py                               (M  – export plugins, tools)
│   ├── plugins/
│   │   ├── __init__.py                           (A)
│   │   └── goodmem/
│   │       ├── __init__.py                       (A)
│   │       ├── client.py                         (A  – HTTP client for Goodmem API)
│   │       └── goodmem_plugin.py                 (A  – chat plugin implementation)
│   ├── tools/
│   │   ├── __init__.py                           (A)
│   │   └── goodmem/
│   │       ├── __init__.py                       (A)
│   │       └── goodmem_tools.py                  (A  – goodmem_save, goodmem_fetch)
│   └── memory/
│       └── goodmem/
│           ├── __init__.py                       (A)
│           └── goodmem_memory_service.py         (A  – BaseMemoryService expansion)
├── tests/unittests/
│   ├── plugins/
│   │   ├── __init__.py                           (A)
│   │   ├── test_goodmem_plugin.py                 (A)
│   │   └── test_goodmem_client.py                 (A)
│   ├── tools/
│   │   ├── __init__.py                           (A)
│   │   └── test_goodmem_tools.py                 (A)
│   └── memory/
│       └── test_goodmem_memory_service.py        (A)
└── contributing/samples/goodmem/
    ├── README.md                                 (A)
    ├── TOOLS.md                                  (A)
    ├── PLUGIN.md                                 (A)
    ├── MEMORY_SERVICE.md                          (A)
    ├── goodmem_tools_for_adk.png                 (A)
    ├── services.py                               (A  – memory service factory for adk web)
    ├── goodmem_tools_demo/
    │   └── agent.py                              (A)
    ├── goodmem_plugin_demo/
    │   └── agent.py                              (A)
    └── goodmem_memory_service_demo/
        └── agent.py                              (A)

Legend: A = added, M = modified.

---

What’s included

Tools (agent-invoked)

Name	Role	When used
GoodmemSaveTool	Wraps goodmem_save	The agent calls it when it wants to store content in Goodmem (e.g. "My favorite color is blue").
GoodmemFetchTool	Wraps goodmem_fetch	The agent calls it when it wants to search/retrieve memories (e.g. "What do I need to do to get into my dream school?").

• goodmem_save: Writes content (text or binary) to a user-scoped Goodmem space with metadata (e.g. user_id, session_id). Space is created or reused per user (adk_tool_{user_id}).
• goodmem_fetch: Runs semantic search over that user’s space and returns the top-k relevant memories (optionally with debug table output).

Plugin (automatic, callbacks)

Name	Role	When triggered
GoodmemChatPlugin	Chat memory for ADK apps	Automatic: on user message → logs user text and supported file attachments to Goodmem; before model → retrieves top-k relevant memories and augments the LLM request; after model → logs the LLM response to Goodmem.

• Uses one Goodmem space per user (adk_chat_{user_id}).
• Filters file attachments by MIME type for Goodmem (e.g. text, PDF, docx); all files still go to the LLM.

Memory Service (based on BaseMemoryService)

Method	Role	When triggered
add_session_to_memory	Write session to memory	Via after_agent_callback (after each turn).
search_memory	Read from memory	Via preload_memory / load_memory tools.

• Stores paired user/model turns as text memories and binary attachments as separate memories.
• Uses one Goodmem space per app+user (adk_memory_{app_name}_{user_id}).
• Uses the shared GoodmemClient from plugins (persistent HTTP connection, multipart binary upload).

---

How to instantiate and wire to an ADK agent

Local development (including before they are merged into an official google-adk-community release):

# Clone the repository (or navigate to your local clone)
cd adk-python-community

# Install the package in editable/development mode
pip install -e .

Tools: all arguments (including optional)

import os
from google.adk.agents import LlmAgent
from google.adk.apps import App
from google.adk_community.tools.goodmem import GoodmemSaveTool
from google.adk_community.tools.goodmem import GoodmemFetchTool

# GoodmemSaveTool – optional: embedder_id, debug
goodmem_save_tool = GoodmemSaveTool(
    base_url=os.getenv("GOODMEM_BASE_URL"),       # required
    api_key=os.getenv("GOODMEM_API_KEY"),         # required
    embedder_id=os.getenv("GOODMEM_EMBEDDER_ID"), # optional; if omitted, first embedder is used
    debug=False,                                   # optional, default False
)

# GoodmemFetchTool – optional: embedder_id, top_k, debug
goodmem_fetch_tool = GoodmemFetchTool(
    base_url=os.getenv("GOODMEM_BASE_URL"),       # required
    api_key=os.getenv("GOODMEM_API_KEY"),         # required
    embedder_id=os.getenv("GOODMEM_EMBEDDER_ID"), # optional
    top_k=5,                                       # optional, default 5 (max 20)
    debug=False,                                   # optional, default False
)

root_agent = LlmAgent(
    model="gemini-2.5-flash",
    name="goodmem_tools_agent",
    description="A helpful assistant.",
    instruction="Answer user questions to the best of your knowledge.",
    tools=[goodmem_save_tool, goodmem_fetch_tool],
)

app = App(name="goodmem_tools_demo", root_agent=root_agent)

Plugin: all arguments (including optional)

import os
from google.adk.agents import LlmAgent
from google.adk.apps import App
from google.adk_community.plugins.goodmem import GoodmemChatPlugin

goodmem_chat_plugin = GoodmemChatPlugin(
    base_url=os.getenv("GOODMEM_BASE_URL"),   # required
    api_key=os.getenv("GOODMEM_API_KEY"),    # required
    name="GoodmemChatPlugin",                 # optional, default "GoodmemChatPlugin"
    embedder_id=os.getenv("EMBEDDER_ID"),     # optional; if omitted, first embedder from API
    top_k=5,                                   # optional, default 5
    debug=False,                               # optional, default False
)

root_agent = LlmAgent(
    model="gemini-2.5-flash",
    name="root_agent",
    description="A helpful assistant for user questions.",
    instruction="Answer user questions to the best of your knowledge",
)

app = App(
    name="goodmem_plugin_demo",
    root_agent=root_agent,
    plugins=[goodmem_chat_plugin],
)

Memory Service: all arguments (including optional)

import os
from google.adk.agents import LlmAgent
from google.adk.apps import App
from google.adk_community.memory.goodmem import GoodmemMemoryService

memory_service = GoodmemMemoryService(
    base_url=os.getenv("GOODMEM_BASE_URL"),
    api_key=os.getenv("GOODMEM_API_KEY"),
    embedder_id=os.getenv("GOODMEM_EMBEDDER_ID"),
    top_k=10,   # Number of memories to retrieve per search, range: 1-100, default 5
    timeout=60.0,      # seconds, default 30.0
    split_turn=False,  # False: one memory per turn (User+LLM); True: two (User, LLM)
)

agent = Agent(
    model="gemini-2.5-flash",
    name="my_agent",
    instruction="You are a helpful assistant with persistent memory.",
    after_agent_callback=save_to_memory,
    tools=[preload_memory, load_memory],
)

runner = Runner(
    app_name="my_app",
    agent=agent,
    memory_service=memory_service,
)

---

Docs and demos

• contributing/samples/goodmem/README.md – Overview of tools, plugin, and memory service.
• contributing/samples/goodmem/TOOLS.md – Setup and usage for tools.
• contributing/samples/goodmem/PLUGIN.md – Setup and usage for the plugin.
• contributing/samples/goodmem/MEMORY_SERVICE.md – Setup and usage for the memory service.
• contributing/samples/goodmem/goodmem_tools_demo/ – Runnable agent with tools.
• contributing/samples/goodmem/goodmem_plugin_demo/ – Runnable agent with plugin.
• contributing/samples/goodmem/goodmem_memory_service_demo/ – Runnable agent with memory service.

[gemini-code-assist]

Summary of Changes

Hello @forrestbao, I'm Gemini Code Assist¹! I'm currently reviewing this pull request and will post my feedback shortly. In the meantime, here's a summary to help you and other reviewers quickly get up to speed!

This pull request significantly enhances the ADK's capabilities by integrating Goodmem.ai for persistent memory management. It provides agents with both explicit tools for saving and fetching information, and an automatic plugin for maintaining conversational context across interactions. This integration aims to improve the agent's ability to recall past information and provide more personalized and coherent responses.

Highlights

• Goodmem Save Tool: Introduced GoodmemSaveTool allowing ADK agents to explicitly store text content in Goodmem.ai, automatically capturing user_id and session_id as metadata.
• Goodmem Fetch Tool: Added GoodmemFetchTool enabling ADK agents to perform semantic searches and retrieve relevant memories from Goodmem.ai, with configurable top_k results.
• Goodmem Chat Plugin: Implemented GoodmemChatPlugin for automatic conversational memory, logging user messages and LLM responses, and augmenting LLM prompts with retrieved relevant chat history.
• Shared Goodmem Client: A dedicated GoodmemClient was created to handle API interactions, ensuring consistent and robust communication with the Goodmem.ai service for both tools and the plugin.
• Comprehensive Testing and Documentation: New unit tests cover the functionality of both the Goodmem tools and the chat plugin, alongside detailed documentation and runnable demo agents to facilitate adoption and usage.

🧠 New Feature in Public Preview: You can now enable Memory to help Gemini Code Assist learn from your team's feedback. This makes future code reviews more consistent and personalized to your project's style. Click here to enable Memory in your admin console.

Using Gemini Code Assist

The full guide for Gemini Code Assist can be found on our documentation page, here are some quick tips.

Invoking Gemini

You can request assistance from Gemini at any point by creating a comment using either /gemini <command> or @gemini-code-assist <command>. Below is a summary of the supported commands on the current page.

Feature	Command	Description
Code Review	/gemini review	Performs a code review for the current pull request in its current state.
Pull Request Summary	/gemini summary	Provides a summary of the current pull request in its current state.
Comment	@gemini-code-assist	Responds in comments when explicitly tagged, both in pull request comments and review comments.
Help	/gemini help	Displays a list of available commands.

Customization

To customize Gemini Code Assist for GitHub experience, repository maintainers can create a configuration file and/or provide a custom code review style guide (such as PEP-8 for Python) by creating and adding files to a .gemini/ folder in the base of the repository. Detailed instructions can be found here.

Limitations & Feedback

Gemini Code Assist may make mistakes. Please leave feedback on any instances where its feedback is incorrect or counter productive. You can react with 👍 and 👎 on @gemini-code-assist comments. If you're interested in giving your feedback about your experience with Gemini Code Assist for Github and other Google products, sign up here.

You can also get AI-powered code generation, chat, as well as code reviews directly in the IDE at no cost with the Gemini Code Assist IDE Extension.

Footnotes

1. Review the Privacy Notices, Generative AI Prohibited Use Policy, Terms of Service, and learn how to configure Gemini Code Assist in GitHub here. Gemini can make mistakes, so double check it and use code with caution. ↩

[gemini-code-assist]

Code Review

This pull request introduces a significant and valuable integration with Goodmem.ai, adding both agent-invoked tools and an automatic chat memory plugin. The code is well-structured and includes comprehensive documentation and tests. However, there are a few critical issues that should be addressed before merging, primarily concerning code duplication, the use of a global variable for state which can cause race conditions, and performance concerns like an N+1 query pattern. I've also included several medium-severity suggestions to improve maintainability, dependency management, and robustness.

[forrestbao]

Summary

• Consolidated Goodmem client into a shared and removed duplicate tool/plugin clients.
• Switched Goodmem HTTP calls to , added batch memory metadata retrieval, and narrowed exception handling.
• Deferred embedder resolution, fixed docs (PLUGIN/TOOLS), and updated tests for new behavior.

[forrestbao]

Summary

• Consolidated Goodmem client into a shared GoodmemClient and removed duplicate tool/plugin clients.
• Switched Goodmem HTTP calls to httpx, added batch memory metadata retrieval, and narrowed exception handling.
• Deferred embedder resolution, fixed docs (PLUGIN/TOOLS), and updated tests for new behavior.

Tests

• uv run pytest tests/unittests/plugins/test_goodmem_plugin.py -v

[forrestbao]

Replied to each Gemini Code Assist comment with the applied fixes.