LatteAI — Fast, Private, Local ☕
LatteAI brings intelligent code completions, chat, and in-editor assistance — all running entirely on your Mac, with no code ever leaving your machine. Choose from three local models that balance speed and quality to suit your hardware:
- Good – Fastest performance, great quality. Works well even on Intel Macs (8 GB+).
- Better – Balanced performance and quality, ideal for Apple Silicon with 16 GB+ RAM.
- Best – Our most advanced model, delivering superior quality for power users with Apple Silicon and 24 GB+ RAM.
Once enabled, LatteAI integrates seamlessly into your editing flow:
- Inline Edits — Make changes directly at the cursor with
⌘K.
- Chat Anywhere — Ask for explanations, refactors, or code generation with
⌘L.
- Context-Aware — LatteAI understands your open files and selections, working with you, not over you.
- Mac-Native — No browser layers. No lag. No privacy worries.
Language Server Protocol — Smarter Editing, Built In 🧠
Espresso now speaks Language Server Protocol (LSP) — bringing modern code intelligence to more languages, right out of the box. No setup required.
- Completions, Hovers & More
Get smart autocompletions, hover information, jump-to-definition, and inline diagnostics for supported languages as you type. - Works Instantly
Espresso ships with built-in LSP support for popular languages. There’s nothing to install or configure — just open a file and go. - Beautifully Native
All LSP features are presented through Espresso’s clean, responsive interface. No clutter. No terminals. No mystery processes.
Note: Espresso uses a curated set of built-in language servers. Connecting to custom or external servers isn’t supported in this release.
And…
- Major under-the-hood performance work keeps the editor feeling buttery smooth, even on large projects.
- Countless smaller improvements and bug fixes.