# ChatOSS

> A native desktop workspace where AI agents write code in your repos, plan tasks on an integrated Kanban board, and run on local Ollama models or cloud providers.

ChatOSS is a native desktop application that positions itself as a Codex alternative for open-source AI, built by the team at pagecow. It combines agentic coding, task management, and multi-model support into a single workspace, and is available via a one-line shell installer or direct download. The core product source code is private; the public GitHub repository at pagecow/chat-oss-releases serves as the release channel for version manifests and downloadable installers.

## What It Is

ChatOSS is a desktop AI coding workspace that lets users run AI agents directly inside their code repositories. Unlike browser-based chat tools, it ships as a native app with an integrated Kanban board for planning and tracking work, a growing suite of built-in apps (such as a PR Reviewer and Migration Bot), and the ability to run many agents in parallel. It is designed to work with both local Ollama models (free, no account required) and cloud-hosted models accessed through a subscription.

## Agentic Coding Workflow

The core workflow centers on agents that operate inside your repositories rather than just generating text in a chat window. Users can assign tasks, move cards on the integrated Kanban board, and have agents implement features, review pull requests, or run migrations — all within the same interface. The homepage describes it as "more than chat — a suite of apps you can extend."

Key workflow capabilities include:
- Agents write code directly in local repositories
- Integrated Kanban board for planning and shipping with a team of agents
- Run multiple agents in parallel on different models
- Switch models mid-conversation without restarting

## Model Support and Local-First Architecture

ChatOSS is built around open-source model support. It integrates with Ollama so that models like Qwen, DeepSeek, GLM, Minimax, Gemma, and NVIDIA Nemotron run entirely on the user's machine — no account, no credit card, and no usage limits. Cloud versions of these models, as well as closed frontier models, are available through paid plans. Users can bring their own API keys for additional cloud access.

## Extensibility and App Suite

Beyond the built-in coding and Kanban features, ChatOSS supports a growing suite of apps that users can extend or build themselves. The homepage highlights examples like a PR Reviewer and a Migration Bot as first-party apps, and the platform is described as extensible so teams can add their own tooling on top of the core workspace.

## Update: v1.9.10

According to the public release channel on GitHub, the latest version is **v1.9.10**, published on 2026-08-18. The release repo was created in July 2026 and uses an automated GitHub Actions workflow to mirror releases from the private source repo to the public channel. The `latest.json` manifest in the repo is polled by installed copies of ChatOSS to detect and download updates. This active release cadence signals ongoing development and regular shipping.

## Features
- Agentic coding inside local repositories
- Integrated Kanban board for task planning and tracking
- Local Ollama model support (free, no account required)
- Cloud model access (open source and closed frontier models)
- Run multiple agents in parallel
- Switch models mid-conversation
- Built-in PR Reviewer app
- Built-in Migration Bot app
- Extensible app suite — build your own apps
- Web search (Starter and above)
- Priority routing (Ultra plan)
- One-line shell installer

## Integrations
Ollama, Qwen, DeepSeek, GLM (Zhipu AI), Minimax, Gemma, NVIDIA Nemotron

## Platforms
WINDOWS, MACOS, LINUX, WEB, API, CLI

## Pricing
Freemium — Free tier available with paid upgrades

## Version
1.9.10

## Links
- Website: https://chatoss.ai
- Documentation: https://github.com/pagecow/chat-oss-releases
- Repository: https://github.com/pagecow/chat-oss-releases
- EveryDev.ai: https://www.everydev.ai/tools/chatoss