InfiniteCode is an open-source coding agent with a Desktop app, terminal TUI/CLI, and model-neutral Rust runtime for private, enterprise, and OpenAI-compatible model environments. Connect DeepSeek, Qwen, Kimi, Anthropic-compatible APIs, local gateways, or your own model endpoint.

English | 简体中文 | 繁體中文 | 日本語 | Русский

Why InfiniteCode · Features · Tested Models · Tested Platforms · Install · Quick Start · Docs

---

Why InfiniteCode

InfiniteCode is for teams that need a coding agent outside a single hosted model ecosystem. It keeps the desktop experience, terminal workflow, model choice, runtime behavior, and workspace execution under your control.

• Bring your own model - Connect OpenAI-compatible Chat Completions, OpenAI-compatible Responses, Anthropic Messages, DeepSeek, Qwen, Kimi, or private model gateways through provider/model bindings.
• Works in private and intranet environments - Run a single local Rust binary, support offline installation paths, and point InfiniteCode at internal endpoints without depending on a hosted agent service.
• One agent across Desktop and terminal - Use the Desktop app for visual onboarding and daily coding, or the CLI/TUI for terminal-native automation, remote shells, and scriptable workflows.
• Built for agent runtime extensibility - MCP servers, reusable skills, local semantic code search, auditable sessions, permissions, and multi-agent flows are runtime features rather than one-off prompts.

Features

• Built-in semantic code search - Runs a local CPU code-embedding model and combines dense retrieval with BM25 keyword matching, reducing code-search context compared with grep/find-only agent.
• Model-neutral provider runtime - Use provider/model bindings for OpenAI-compatible, Anthropic-compatible, DeepSeek, Qwen, Kimi, GLM, MiniMax, Xiaomi MiMo, OpenRouter, or local endpoints.
• MCP support - Connect external tools and context through Model Context Protocol servers.
• Skill support - Package repeatable workflows, instructions, scripts, and references as reusable Agent Skills.
• Long-running task support - Let InfiniteCode manage context automatically across multi-turn work instead of losing the thread as tasks grow.
• Multi-agent support - Split work across specialized agents while keeping coordination visible in the session.
• Plan Mode - Break larger tasks into clear multi-step plans before implementation starts.
• Parallel tool calls - Run multiple independent tools in parallel so models spend less time waiting and more time making progress.
• Permissioned tool execution - Review sensitive tool calls before they touch your workspace.
• Auditable sessions - Keep model output, tool calls, approvals, token usage, and session history inspectable and resumable.
• Cost and context visibility - Show input/output tokens, cached tokens, and context-window usage where providers expose them.
• Lightweight Rust runtime - Built in Rust with low memory overhead and a compact local runtime.

Tested Models

InfiniteCode's built-in model catalog includes tested model definitions for Qwen, Kimi, MiniMax, GLM, and DeepSeek. Provider endpoints remain configurable through provider/model bindings.

Tested Platforms

InfiniteCode has been tested on macOS, Linux, Windows, and Kylin OS.

For Chinese Enterprise Users

Kylin OS coverage is called out because domestic operating systems are often part of real deployment requirements in Chinese enterprise environments. HarmonyOS support is on the roadmap; contributors with HarmonyOS devices are welcome to build, test, and publish releases for that platform.

Installation

InfiniteCode can be installed in two forms. Pick the Desktop app for a graphical coding agent workspace, the terminal-native TUI/CLI for shell-first development, or install both on the same machine.

Option 1: Desktop App

Start here if you want the graphical InfiniteCode experience. Download the latest InfiniteCode Desktop package from GitHub Releases, then choose the asset that matches your operating system and architecture:

• macOS - download the infinitecode-desktop-...-mac-... .dmg or .zip asset.
• Windows - download the infinitecode-desktop-...-windows-... .exe asset.
• Linux - download the infinitecode-desktop-...-linux-... .AppImage, .deb, or .rpm asset.

If macOS reports that InfiniteCode.app is damaged and cannot be opened, this is expected. Current macOS Desktop builds are unsigned, so after installing, run the following command so macOS can launch the app:

sudo xattr -dr com.apple.quarantine /Applications/InfiniteCode.app

Option 2: TUI / CLI

Install the terminal-native infinitecode command if you prefer the TUI, want shell automation, or want to use InfiniteCode alongside the Desktop app.

Linux / macOS:

curl -fsSL https://raw.githubusercontent.com/AFS-Agentics/InfiniteCode/main/install.sh | sh

Windows:

irm 'https://raw.githubusercontent.com/AFS-Agentics/InfiniteCode/main/install.ps1' | iex

The online installer places infinitecode under the InfiniteCode home directory, installs the rg sidecar used for fast repository search, and supports optional setup for the local model used by code_search.

Optional: preinstall the local code_search model

Use this only if you want the Hugging Face model downloaded during installation.

Linux / macOS:

curl -fsSL https://raw.githubusercontent.com/AFS-Agentics/InfiniteCode/main/install.sh | sh -s -- --install-code-search-model

Windows:

$env:INFINITECODE_INSTALL_CODE_SEARCH_MODEL = "1"; irm 'https://raw.githubusercontent.com/AFS-Agentics/InfiniteCode/main/install.ps1' | iex

Upgrade an existing installation to the latest release:

infinitecode upgrade

The upgrade command runs the same platform installer, and the installer prints the version transition, for example Version: v0.1.12 -> v0.1.15.

For air-gapped or intranet installs, see Offline Installation.

Quick Start

Configure a provider, open a repository, and start the TUI:

cd /path/to/your/repo
infinitecode onboard

Useful commands:

infinitecode                         # start the interactive TUI in the current repo
infinitecode resume <session-id>

Configuration

infinitecode onboard is the recommended setup path. For manual config.toml paths, provider/model binding fields, and custom model catalog examples, see Configuration.

Docs

• Offline Installation
• Configuration

FAQ

What is the project status?

InfiniteCode is pre-1.0 and actively developed. It is ready for local evaluation, experiments, and contributor use; public APIs and configuration may still change.

What models are supported?

Built-in model metadata currently covers Qwen, Kimi, MiniMax, GLM, and DeepSeek families. Any model endpoint that supports OpenAI-compatible Chat Completions, OpenAI-compatible Responses, or the Anthropic Messages API can be connected through provider/model bindings.

Should I use the Desktop app or the TUI/CLI?

Use the Desktop app when you want visual onboarding, session browsing, and a graphical coding workspace. Use the TUI/CLI when you want terminal-native automation, remote shell workflows, or a coding agent that stays inside your existing command-line setup. Both surfaces target the same local InfiniteCode runtime.

Contributing

Contributions are welcome while the project is still early:

• Architecture feedback on the client/server runtime, provider layer, safety model, and TUI.
• Documentation and translations.
• Provider, model, and wire API coverage.
• Focused fixes with validation commands and regression tests.

Open an issue or pull request to discuss changes.

Star History

License

This project is licensed under the MIT License.

---

If you find InfiniteCode useful, please consider giving it a star.