{"slug": "memorymate-turning-friend-conversations-into-memories", "title": "MemoryMate — Turning Friend Conversations Into Memories", "summary": "A developer built MemoryMate, an AI-powered personal memory assistant that ingests conversations, notes, screenshots, PDFs, and transcripts and answers natural-language questions about them with source context. The system combines Gemma for memory extraction and grounded responses with Tiger Data (PostgreSQL plus vector search) for hybrid semantic and structured retrieval, and includes a \"Friend Mode\" that builds per-person profiles of preferences, dates, and interests. The backend is built with FastAPI and deployed on Render, with a React frontend, and the code is available on GitHub.", "body_md": "*This is a submission for the [Hacktoberfest Weekend Challenge: Build for a Friend](https://dev.to/challenges/hacktoberfest-weekend-2026-10-01)*\n\n🧠 MemoryMate — An AI built to help a friend remember the little things.\n\nFriends talk about hundreds of things — ideas, plans, birthdays, recommendations, and random conversations.\n\nBut sometimes we forget the details that matter:\n\n“What did I say I wanted?”\n\n“When did we talk about that?”\n\n“What was that idea we discussed?”\n\nI built MemoryMate to help a friend find those memories again — without searching through hundreds of messages.\n\nMemoryMate is an AI-powered personal memory assistant that helps you remember important things from your conversations, notes, screenshots, PDFs, transcripts, and plans.\n\nInstead of scrolling through hundreds of messages, you can simply ask:\n\n💬 “What were the startup ideas Rahul and I discussed?”\n\n💬 “When did I say my interview was?”\n\n💬 “What movie did my friend recommend?”\n\n💬 “What does Rahul like?”\n\nMemoryMate searches your personal memories, finds the relevant context, and gives you an answer with the original source so you can verify where the information came from.\n\n✨ Friend Mode\n\nThe most interesting part is Friend Mode.\n\nMemoryMate can build a lightweight memory profile for each friend:\n\nRahul\n\n❤️ Football · Gaming\n\n🚀 College Startup\n\n🎧 Looking for headphones\n\n🎂 Birthday · May 18\n\nAnd this enables one of my favorite features:\n\n💡 “You mentioned this before…”\n\nImagine saying:\n\n“I need a birthday gift for Rahul.”\n\nMemoryMate can look through previous conversations and respond:\n\nYou mentioned this before: Rahul has been looking for headphones and has recently been talking about gaming. He also likes football.\n\nInstead of just answering questions, MemoryMate helps you remember the people you care about.\n\n🚀 Render Link:  [https://memorymate-lr5k.onrender.com/](https://memorymate-lr5k.onrender.com/)\n\n💻 GitHub: [https://github.com/yogesh147/memorymate](https://github.com/yogesh147/memorymate)\n\nThe repository contains the frontend, backend, AI pipeline, memory retrieval system, and supporting infrastructure used to build MemoryMate.\n\nMemoryMate is built around open-weight AI, semantic search, structured memory, and retrieval-augmented generation (RAG).\n\n```\n      📄 Conversations\n      📝 Notes\n      🖼️ Screenshots\n      📑 PDFs\n      🎙️ Transcripts\n             │\n    Ingestion Pipeline\n             │\n  Extract • Chunk • Structure\n             │\n    Memory Extraction\n             │\n People • Dates •  Facts • Preferences\n             │\n         Tiger Data\n  PostgreSQL + Vector Search\n             |\n       Hybrid Retrieval\n             │\n           Gemma\n             │\n    Answer + Source Context\n```\n\n🧠 Gemma — The Intelligence\n\nUse Gemma as the open-weight language model powering the AI layer.\n\nGemma is used for:\n\nExtracting useful memories from conversations\n\nIdentifying people, dates, preferences, and events\n\nUnderstanding natural-language questions\n\nReasoning over retrieved memories\n\nGenerating grounded responses\n\nBuilding Friend Mode profiles\n\nRather than treating the LLM as a simple chatbot, Use as part of a memory pipeline where retrieved information provides the context for each answer.\n\n🗄️ Tiger Data — The Memory\n\nMemoryMate uses Tiger Data as the database and vector-search layer.\n\nA memory contains both semantic and structured information:\n\nMemory\n\n├── Content\n\n├── Embedding\n\n├── Person\n\n├── Timestamp\n\n├── Conversation\n\n├── Source\n\n├── Confidence\n\n└── Metadata\n\nThis allows the application to combine:\n\nSemantic search\n\n“What did Rahul say about his startup?”\n\nwith:\n\nStructured search\n\n“Find things Rahul said about his startup in August.”\n\nThat combination makes personal-memory retrieval much more useful than relying on vector similarity alone.\n\n🚀 Render — The Application\n\nUse Render to deploy the MemoryMate backend and expose the application through a production API.\n\nThe backend is built with FastAPI, while the frontend is built with React.\n\nThis gives me a simple architecture:\n\nReact\n\n  ↓\n\nFastAPI\n\n  ↓\n\nMemory / Retrieval Layer\n\n  ↓\n\nTiger Data + Gemma\n\n🔐 Privacy & Local AI\n\nPersonal memories are sensitive.\n\nThat's why I designed MemoryMate around open-weight models and local inference.\n\nThe architecture can be adapted to run AI inference locally, giving users more control over where their memories and personal information are processed.\n\nThe goal isn't just to make AI remember more.\n\nIt's to explore how AI can remember more responsibly.\n\nOpen innovation allowed me to combine open-weight AI, vector search, local inference, and open-source tools into a privacy-focused personal AI.\n\nFor a product handling personal memories, this control matters.\n\nInstead of relying entirely on a closed API, developers can customize how memories are extracted, stored, searched, and processed.\n\nMemoryMate shows how open AI can become more than a chatbot — it can become a personal memory layer.\n\nI'm entering MemoryMate for:\n\n🟣 Best Use of Render - powers the deployed MemoryMate backend and API.\n\n🟢 Best Use of Gemma - powers memory extraction, reasoning, Friend Mode and grounded responses.", "url": "https://wpnews.pro/news/memorymate-turning-friend-conversations-into-memories", "canonical_source": "https://dev.to/yogesh147/memorymate-turning-friend-conversations-into-memories-1k0h", "published_at": "2026-10-04 10:05:02+00:00", "updated_at": "2026-10-04 10:12:24.861249+00:00", "lang": "en", "topics": ["artificial-intelligence", "large-language-models", "ai-products", "ai-tools", "generative-ai"], "entities": ["MemoryMate", "Gemma", "Tiger Data", "Render", "FastAPI", "React", "GitHub", "Hacktoberfest"], "also_reported_by": [], "alternates": {"html": "https://wpnews.pro/news/memorymate-turning-friend-conversations-into-memories", "markdown": "https://wpnews.pro/news/memorymate-turning-friend-conversations-into-memories.md", "text": "https://wpnews.pro/news/memorymate-turning-friend-conversations-into-memories.txt", "jsonld": "https://wpnews.pro/news/memorymate-turning-friend-conversations-into-memories.jsonld"}}