Free, open-source desktop app that turns your journals, chats and data exports into a searchable timeline of how your thinking evolved. Local by default.
A free, open-source disk-space visualizer. Squarified treemap, duplicate hunting, disk-full forecasting, and cleanup that only ever moves files to the Trash.
Ayakaleaf Pro is a free, open-source, self-hosted Overleaf alternative: Overleaf Community Edition with almost all Server Pro features — track changes, project history, admin panel, SSO, Git-bridge, Z
Privacy-first native app that builds a searchable, visual timeline of how your thinking evolved. Tauri 2.0 + Rust + React.
翻译准备中
取自仓库 README
README
正在加载完整 README。 以下为已保存的摘录。
A searchable, visual timeline of how your thinking evolved.
MemryLab is a privacy-first native desktop application that ingests your entire digital footprint -- journal entries, chat exports, social media archives, notes, emails -- and constructs a searchable, visual timeline of how your thinking, interests, and emotional patterns have evolved over time. It supports 30+ import sources with automatic format detection, runs an 8-stage AI analysis pipeline, and provides keyword, semantic, and hybrid search alongside a RAG-powered chat interface.
MemryLab is a privacy-first native desktop application that ingests your entire digital footprint -- journal entries, chat exports, social media archives, notes, emails -- and constructs a searchable, visual timeline of how your thinking, interests, and emotional patterns have evolved over time. It supports 30+ import sources with automatic format detection, runs an 8-stage AI analysis pipeline, and provides keyword, semantic, and hybrid search alongside a RAG-powered chat interface. Everything runs locally in a single SQLite database with zero telemetry, zero cloud dependency, and a 4.3 MB installer.
No external services required -- fully functional with Ollama alone
AI Providers
9 providers with one-click preset setup, 8 with free tiers
Separate LLM and embedding providers -- use a cloud LLM for analysis + local Ollama for private embeddings
Universal OpenAI-compatible adapter -- works with any endpoint that speaks the OpenAI API format
Runtime provider switching -- change models without restarting
Usage logging for token tracking
Quick Start
Option A: Download Installer
Download the latest release from the Releases page.
Platform
Installer
Size
Windows
.exe (NSIS)
~4.3 MB
macOS
.dmg
Coming soon
Linux
.AppImage / .deb
Coming soon
Option B: Build from Source
Prerequisites:Rust 1.70+, Node.js 18+, Ollama (optional, for local LLM)
git clone https://github.com/laadtushar/MemryLab.git
cd MemPalace
npm install
# (Optional) Pull local models — see Recommended Models table above
ollama pull nomic-embed-text
ollama pull qwen2.5:14b-instruct-q5_K_M # or llama3.1:8b for less VRAM
# Development
cargo tauri dev
# Production build
cargo tauri build
Recommended Models (Ollama / Local)
Pick models based on your hardware. Embedding model is required for search; LLM is required for analysis.
VRAM
LLM Model
Embedding Model
Speed
4 GB
llama3.2:3b
nomic-embed-text
~40 tok/s
8 GB
llama3.1:8b
nomic-embed-text
~35 tok/s
12 GB
qwen2.5:14b-instruct-q5_K_M
nomic-embed-text
~25 tok/s
16 GB+
qwen2.5:32b-instruct-q4_K_M
nomic-embed-text
~15 tok/s
CPU only
llama3.2:3b
nomic-embed-text
~5 tok/s
# Example for 12 GB VRAM (e.g., RTX 4070/5070 Ti)
ollama pull nomic-embed-text
ollama pull qwen2.5:14b-instruct-q5_K_M
Tip: Larger models produce better belief extraction and contradiction detection. If you have the VRAM, use the biggest model that fits. For cloud alternatives with no hardware requirements, configure Gemini Flash (nearly free) or Groq (free tier) in Settings.
First Run
Open the app and go to Settings to pick an AI provider (Ollama, Groq, Gemini, etc.)
Click Import in the sidebar -- select a source type or drop a ZIP/folder
Run Analysis from the Insights view to extract themes, beliefs, and entities
Search, Ask, or explore the Timeline, Graph, and Evolution views
Any OpenAI-compatible endpoint can also be configured manually with a custom base URL.
Architecture
MemryLab follows a strict hexagonal (ports and adapters) architecture. The core domain has zero knowledge of specific databases, LLM providers, or UI frameworks.
flowchart LR
A[Drop File/ZIP] --> B[Auto-Detect Source]
B --> C[Platform Adapter]
C --> D[Dedup by SHA-256]
D --> E[Normalize Unicode]
E --> F[Chunk 512 tokens]
F --> G[Generate Embeddings]
G --> H[Store + FTS5 Index]
H --> I[Auto-Run Analysis]
style A fill:#7c3aed,stroke:#a78bfa,color:white
style I fill:#22c55e,stroke:#4ade80,color:white
Analysis Pipeline
flowchart LR
T[Themes] --> S[Sentiment]
S --> B[Beliefs]
B --> E[Entities]
E --> I[Insights]
I --> C[Contradictions]
C --> EV[Evolution]
EV --> N[Narratives]
style T fill:#7c3aed,stroke:#a78bfa,color:white
style N fill:#22c55e,stroke:#4ade80,color:white
The hexagonal architecture means adding a new source adapter, storage backend, or analysis stage only requires implementing one trait interface -- no changes to business logic.
Fork the repository
Create a feature branch (git checkout -b feature/amazing-feature)
Commit your changes (git commit -m 'feat: add amazing feature')
Push to the branch (git push origin feature/amazing-feature)
