Skip to content

memsync-chatbot

Directory actions

More options

Directory actions

More options

Latest commit

 

History

History

memsync-chatbot

Folders and files

NameName
Last commit message
Last commit date

parent directory

..
README.md
README.md
 
 
chatbot.py
chatbot.py
 
 
requirements.txt
requirements.txt
 
 

README.md

MemSync Chatbot 🧠

A personalized AI chatbot that remembers users across sessions using MemSync's semantic memory layer and OpenGradient's verifiable LLM.

Architecture

User message
     │
     ▼
MemSync search ──── POST /memories/search
     │  (relevant memories + user_bio)
     ▼
OpenGradient LLM ─── TEE-verified inference
     │  (enriched system prompt + memory context)
     ▼
Response + payment_hash
     │
     ▼
MemSync store ──── POST /memories
     │  (auto-extracts new facts from conversation)
     ▼
Persistent memory for next session

Setup

cd boilerplates/memsync-chatbot
pip install -r requirements.txt
cp ../../.env.example .env

Edit .env:

OG_PRIVATE_KEY=0xYOUR_TESTNET_KEY
MEMSYNC_API_KEY=your_memsync_key  # from app.memsync.ai/dashboard/api-keys

Run

python chatbot.py

Example Conversation

Session 1:

You: Hi! I'm a Python developer working on a DeFi protocol.
Bot: Great to meet you! What kind of DeFi protocol are you building?

You: An automated market maker with AI-driven fee adjustment.
Bot: That's a fascinating use case — AI-driven AMM fees could significantly
     optimize capital efficiency...

Session 2 (next day):

You: I'm having trouble with the fee calculation logic.
Bot: Happy to help! Given your work on the AI-driven AMM fee adjustment
     system, are you running into issues with the fee curve parameters
     or the ML model integration?
     # ^ Bot remembers from Session 1!

Memory Types

Type Description Example
semantic Lasting facts about the user "User is a Python developer"
episodic Time-bound events "User mentioned fee calculation issue on March 14"

Commands (Interactive REPL)

Command Action
profile Show your current MemSync profile + insights
debug Toggle memory retrieval debug output
quit Exit (memories persist)

Programmatic Usage

from chatbot import MemoryChatbot
import asyncio

bot = MemoryChatbot(agent_id="my-app", show_memory_debug=True)

async def main():
    response, payment_hash = await bot.chat(
        "What tech stack should I use for my DeFi project?"
    )
    print(f"Bot: {response}")
    print(f"Proof: {payment_hash}")

    # View stored profile
    profile = bot.get_profile()
    print(profile["user_bio"])

asyncio.run(main())

Resources