Skip to content
CodeLocal.
Open dashboard
Project intelligence. Local control.

CodeLocal

A shared intelligence layer for the way you build with AI.

Connect your AI to one Project Brain. Keep execution on your machine.

Explore CodeLocal See how it works
Your workspace. Your permission.Context / Execution / Memory
Built to connect with your AI
ChatGPTCodexClaudeCompatible MCP clients
01 / The missing layer

Better AI. Not another fresh start.

Models evolve. Tools change. Your project should keep learning.

PROJECT BRAIN

Context that carries forward.

Project knowledge, reusable skills and verified experience give your AI a foundation beyond the current conversation.

LOCAL EXECUTION

Your machine. Your control.

Work with the tools and files on your own device, within explicit workspace and approval boundaries.

MODEL-NEUTRAL

Change models. Keep context.

Connect compatible AI clients through MCP. Your project intelligence does not have to start over with each tool.

02 / One connected system

Intelligence shared.
Boundaries preserved.

AI clients, cloud coordination and local execution. Connected through one project-aware system.

The CodeLocal modelArchitecture
One connected system. Your machine at the center of execution.AI clients connect by MCP to CodeLocal. CodeLocal routes approved requests to your local machine. Project Brain provides sanitized durable knowledge. Cloud and local execution have separate privacy boundaries.
Cloud: identity, routing, knowledge
Local: your privacy boundary
MCPRPCContext01 / AI
Different AIs. One entry.ChatGPT · Claude · Codex
02 / Control layerCodeLocal03 / Your machine
Local power. Your rules.Files · Git · Terminal · Browser
04 / PROJECT BRAIN
Intelligence that stays.Decisions · Knowledge · Verified experience
System model, not live activity
02 / Connect through CodeLocal

CodeLocal resolves the project, supplies bounded context and routes requests. Workspace authorization and approval policy determine what can run.

Scoped by design.

Local tools operate on your machine. Cloud services coordinate identity, routing and sanitized project knowledge.

Security model
03 / Product

One place to see the whole project.

Keep your connected workspaces, project knowledge and access controls in view.

Open dashboard
04 / Get connected

Your machine. Three steps to connected.

Connection guide
  1. 01

    Install the runtime

    On the machine where your code lives.

    npm install -g codelocal@latest
  2. 02

    Connect your workspace

    Pair your device and choose your project.

    codelocal .
  3. 03

    Bring your AI

    Connect a compatible client through your MCP endpoint.

    One project. Shared context.
On the roadmap

CodeLocal Desktop

macOS · Windows · Linux

In development
Build on what you know

The next model will change.
Your project stays yours.

Get started Talk to us