Open source coding agent

BOWOW

A desktop coding agent that reads, writes, and diffs your project files directly on disk — connect any model, describe what to build, and let it run.

Download Now View on GitHub
Bowow demo
Capabilities

Everything the agent brings to the table

Multi-model, multi-session, and built to keep your project files honest.

Multi-Model Support

Gemini, OpenAI, Ollama & more

Core

Split-Screen Mode

4-pane view (F10)

UI

Live File Editing

Direct disk read/write/diff

IO

Checkpoint System

Rollback file changes

Safe

Context Management

Auto compaction & pruning

AI

Error Auto-Retry

Exponential backoff

Stable

Multi-Model Support

Gemini, OpenAI, Ollama & more

Core

Split-Screen Mode

4-pane view (F10)

UI

Live File Editing

Direct disk read/write/diff

IO

Checkpoint System

Rollback file changes

Safe

Context Management

Auto compaction & pruning

AI

Error Auto-Retry

Exponential backoff

Stable

MCP Tools

Experimental integration

Ext

System Prompt

Define AI behavior

Config

User Prompts

Manage multiple prompts

User

Terminal Commands

Enhanced control

Shell

Auto-Update System

Effortless updates

Sync

MCP Tools

Experimental integration

Ext

System Prompt

Define AI behavior

Config

User Prompts

Manage multiple prompts

User

Terminal Commands

Enhanced control

Shell

Auto-Update System

Effortless updates

Sync
Preview

Inside the app

Bowow desktop preview
Bowow desktop preview 2
Get started

Usage

01
Configure your provider

Open Settings, add your AI provider and API key, then click Done.

02
Connect a model

Gemini, OpenAI, Ollama, or any supported backend.

03
Open a project

Use /new to open a project folder.

04
Describe what to build

The agent generates files, edits code, and runs commands.

05
Go multi-session

Press F10 to toggle split-screen mode.