Organize Projects.
Empower Local AI.
Zero Telemetry.
A native Linux desktop client and web workstation built for Ollama, cloud LLMs, and swappable live coding harnesses — with persistent agent memory, project workspaces, and strict local privacy.
Engineered for AI Autonomy & Developer Privacy
Everything you expect from modern AI developer tooling — without corporate telemetry or closed-source lock-in.
100% Free & Offline-Ready
Connect directly to Ollama to run cutting-edge models (Llama 3.2, DeepSeek-R1, Qwen 2.5-Coder, Mistral, Phi-4) on your local GPU/NPU with zero monthly subscriptions.
Live Coding Workspace Engines
Swappable Pi and DeepSeek coding harnesses with live diff reviews, sandboxed terminal execution, approval prompts, and read-only safety modes.
Workspaces & Project Memory
Organize chats into dedicated project workspaces. Persistent RAG vector memory (nomic-embed-text) automatically remembers your coding style and architecture.
Universal AI Model Routing
Plug in OpenRouter free models, high-speed 9Router proxy, OpenAI (GPT-4o, o1, o3-mini), Anthropic Claude, or local endpoints like LM Studio and vLLM.
Git Status & Safety Creator
Real-time branch monitoring in the workspace bar with warning badges on production branches and 1-click safety branch creation before code generation.
Themes & Native Linux Ergonomics
Native Linux client built with Tauri 2.0 and Rust. Includes Vanaila Origin and Catppuccin color schemes (Teal, Rose, Blue, Green, Peach) with light & dark modes.
Dedicated Workspaces with Independent Memory
Never mix unrelated client chats or research notes again. VanailaChat auto-organizes conversations by project workspace, linking each to a local codebase folder while maintaining isolated prompts, custom instructions, and long-term semantic memory.
- ✓Isolated workspace memory and custom system instructions per project
- ✓Linked workspace directory with real-time Git status and safety branching
- ✓Semantic vector search powered by nomic-embed-text for instant recall
- ✓Accurate live token meters tracking prompt and completion tokens
Interactive Coding Harnesses with Live Diff Inspection
Harness the power of AI coding agents directly on your machine. Choose between the lightweight Pi Coding Harness or the multi-file DeepSeek Harness. Every tool execution is transparent: inspect touched files, review diffs line by line, and approve actions with a keystroke.
- ✓Swappable Pi and DeepSeek harnesses using your configured AI provider
- ✓Dedicated right activity drawer with live terminal feeds and approval prompts
- ✓Side-by-side git diff review before accepting code changes
- ✓Read-only safety switch to allow code research without accidental edits
Get Started in Seconds
Download the native Linux installer for your distribution or launch the automated cross-platform web edition.
| Platform / Distro | Package | Quick Install Command | Action |
|---|---|---|---|
| Ubuntu / Debian / Mint / Pop!_OS | .deb (x86_64) | sudo apt install ./vanaila-chat_0.3.2_amd64.deb | Download .deb → |
| Fedora / RHEL / openSUSE | .rpm (x86_64) | sudo dnf install ./vanaila-chat-0.3.2-1.x86_64.rpm | Download .rpm → |
| Universal Linux (Any Distro) | .AppImage | chmod +x *.AppImage && ./*.AppImage | Download .AppImage → |
| Arch Linux / Manjaro | AUR Package | yay -S vanaila-chat-bin | View on AUR → |
| 1-Click Web App (Linux / Mac / Windows) | Automated Script | ./start.sh (Linux/Mac) or start.bat (Win) | Clone & Run → |
Made with Love for the Indonesian AI Community
An open-source, privacy-first AI workstation built to empower students, independent creators, and senior software engineers to build with modern AI — completely free and sovereign.
Join the Community on GitHub→Reclaim Your AI Sovereignty Today
Run your models locally, preserve your code confidentiality, and experience a developer-first AI workspace.