Features

One editor.
Every superpower.

CodeGrex pairs the editor you know with autonomous agents, local‑first privacy, and completions that learn your codebase — all designed to keep you in flow.

Download CodeGrexView pricing
Autocomplete

Magically accurate completions.

A specialized Tab model predicts your next edit with striking speed and precision. Accept with a single keystroke and never break your flow.

  • Sub‑100ms latency with local models
  • Context‑aware multi‑line completions
  • Smart Tab acceptance with partial accept
  • Ghost‑text preview before accepting
completion.ts
async function fetchUserData(userId) {
const res = await api.get(`/users/$${userId}`);
return res.data;
}
// Press Tab to accept suggestion ⇥
agent
✓Reading project structure
✓Creating components/Header.tsx
✓Updating App.tsx imports
4Running npm install tailwindcss
5Updating tailwind.config.js
Agent mode

Autonomous AI agents.

Hand off complex, multi‑file tasks. The agent reads your codebase, edits across files, runs terminal commands, and fixes errors automatically.

  • Multi‑file editing and creation
  • Execute terminal commands in a sandbox
  • Automatic error detection and fixing
  • Git operations and commit management
Inline edit · ⌘K

Edit code in natural language.

Select any code and press ⌘K to describe your change. Review a clean diff preview, then accept or reject — hunk by hunk.

  • Select and describe changes
  • Visual diff preview
  • Iterate until it's perfect
⌘K — inline edit
function add(a, b) {
function add(a: number, b: number): number {
return a + b;
}
“Add TypeScript types to this function”
Chat & context

AI that knows your code.

Reference files, folders, and symbols with @ mentions. The agent understands your whole codebase.

@file

File context

@file:src/utils/auth.ts
@folder

Folder context

@folder:src/components
@codebase

Codebase search

@codebase: auth logic
@docs

Documentation

@docs: API reference
@web

Web search

@web: React 19 features
@terminal

Terminal output

@terminal: last error
Local AI

Privacy‑first local models.

Run AI completely offline on your own hardware — your code never leaves your machine. Perfect for air‑gapped environments and sensitive code.

  • 100% offline operation
  • GPU‑accelerated inference
  • One‑command setup with Ollama
  • Any OpenAI‑compatible server (LM Studio, llama.cpp, vLLM)
Set up local models
OllamaFast

Qwen2.5‑Coder

7B params

OllamaMedium

DeepSeek Coder

6.7B params

OllamaMedium

CodeLlama

7B params

OpenAI‑APIFast

Starcoder2

3B params

Cloud models

Access the best AI models.

Use managed credits or bring your own keys. One interface, every frontier model.

Anthropic

  • Claude Opus 4.8
  • Claude Opus 4.7
  • Claude Sonnet 4.6
  • Claude Haiku 4.6

OpenAI

  • GPT‑5.5
  • GPT‑5.4
  • GPT‑5.3
  • GPT‑5.2

Google

  • Gemini 3.1 Pro
  • Gemini 3.5 Flash
  • Gemini 3 Flash
  • Gemini 2.5 Pro

Moonshot

  • Kimi K2.6
  • Kimi K2.5
  • Kimi K2 Thinking Turbo
Bring your own key (BYOK) for every provider

And so much more.

Everything you need for productive AI‑assisted coding.

Memories

AI remembers your codebase patterns and preferences.

Smart diffs

See exactly what changes before you accept.

MCP support

Connect to Figma, Slack, and other tools.

Codebase indexing

Semantic search across your entire project.

Web search

Pull current information into your prompts.

Extensions

Full compatibility with the extensions you already use.

Continue work

Resume exactly where you left off.

Auto lint fix

Automatically resolve linter errors.

Customizable

Themes, keybindings, and deep settings.

Ready to feel the difference?

Download CodeGrex for free and start coding with AI superpowers.

Download CodeGrex