Wxw-Gu/TraceMemo

TraceMemo（迹忆） – 本地优先的微信数据、AI 分析与自动化工作台

What it is – TraceMemo is a desktop application (Windows x64, macOS Apple Silicon & Intel) that lets you treat your WeChat chat history like a searchable, query‑able knowledge base. It reads the local WeChat database, builds indexes on text and on OCR‑extracted text from images, and optionally layers an LLM‑powered “AI Search” on top. The app also bundles a set of automation features such as daily group‑chat summaries, member‑exit monitoring, and text‑to‑speech generation.

Core capabilities

• Archive & search – Browse conversations, filter by keyword, nickname, wxid, or custom remarks.
• AI Search / “Ask WeChat” – Pose natural‑language questions; the LLM returns matching messages together with their source location.
• Local knowledge base – All indexes live on the user’s machine, ensuring privacy and fast cross‑conversation queries.
• Image‑text OCR – On‑device OCR extracts text from screenshots, announcements, price‑list images, making that content searchable without uploading anything.
• Group‑chat reports – Generate daily/weekly HTML or PNG summaries of a group’s discussion, including speaker statistics.
• Member‑exit monitoring – Snapshot‑based tracking of who leaves a group, with a history of events.
• Automation – Chain the above steps into scheduled jobs (e.g., auto‑send a daily report or a leave‑notification) via the built‑in “Automation” engine.
• Text‑to‑speech – Convert any message to a voice file and send it back into the chat.
• Agent Hub & API – Interact with TraceMemo from within WeChat (local agent) or from external agents (e.g., Codex) through a local HTTP API.

How it works – The app first connects to the local WeChat SQLite database (compatible with WeChat 4.x) and obtains the encryption key needed to decrypt messages. It then builds searchable indexes for plain text and OCR‑derived text. When an AI model is configured (OpenAI, Claude, etc.), queries are sent to the model, which returns a natural‑language answer plus references to the underlying messages.

Why it matters – For power users, researchers, or teams that rely heavily on WeChat for coordination, TraceMemo turns an otherwise opaque chat log into a first‑class information resource while keeping all data on the user’s device. The combination of local OCR, a persistent knowledge base, and optional LLM‑driven retrieval bridges the gap between raw chat archives and actionable insights.

Getting started – Download the appropriate installer from the GitHub Releases page, point the app at your WeChat data folder, and optionally add an AI model under Settings → AI Model. The documentation includes a step‑by‑step “first‑time use” guide, detailed user manuals, and contribution guidelines.

---

TL;DR – TraceMemo is a privacy‑first desktop tool that indexes your WeChat history, lets you search it with keywords or LLM‑powered natural language, generates automated group reports, monitors member changes, and offers extensibility via a local API and agents.