# o4

> Documentation for o4, the coding agent for your terminal.

- [Introduction](https://docs.open4rena.ai/index.md): What o4 is, what it can do, and how it works.
- [Install o4](https://docs.open4rena.ai/installation.md): Install o4 on macOS or Linux with the install script, Homebrew, or a release archive.
- [Quickstart](https://docs.open4rena.ai/quickstart.md): Connect a model and finish your first task with o4.
- [The interface](https://docs.open4rena.ai/guides/interface.md): How the o4 terminal interface is laid out, and how its input modes work.
- [Writing prompts](https://docs.open4rena.ai/guides/prompting.md): Type prompts, point o4 at files, attach images, and send messages while it works.
- [Sessions](https://docs.open4rena.ai/guides/sessions.md): Resume, fork, rename, export, and delete saved conversations.
- [Print mode and scripting](https://docs.open4rena.ai/guides/print-mode.md): Run o4 once without the interface, for scripts, pipelines, and CI.
- [Context and cost](https://docs.open4rena.ai/guides/context-and-cost.md): See how much of the context window and budget a session uses, and compact it.
- [Git and code review](https://docs.open4rena.ai/guides/git-and-review.md): Review the changes o4 made, commit them, and ask for a code review.
- [Undo and rewind](https://docs.open4rena.ai/guides/checkpoints.md): Undo o4's last edits, or rewind files and the conversation to an earlier turn.
- [Watch mode](https://docs.open4rena.ai/guides/watch.md): Run your project's check whenever files change, and let o4 fix failures.
- [Plan mode](https://docs.open4rena.ai/guides/plan-mode.md): Let the model research and plan before it changes any files.
- [Goals](https://docs.open4rena.ai/guides/goals.md): Give the model one objective to keep working on across turns.
- [Todos and background tasks](https://docs.open4rena.ai/guides/tasks-and-todos.md): Track the model's todo list, and run shell commands and prompts in the background.
- [Subagents](https://docs.open4rena.ai/guides/subagents.md): Let the model hand work to subagents that run in their own context.
- [Campaigns](https://docs.open4rena.ai/guides/campaigns.md): Split one objective across several agents that work in parallel.
- [Memory](https://docs.open4rena.ai/guides/memory.md): What o4 remembers between sessions, where it's stored, and how to clear it.
- [Code intelligence](https://docs.open4rena.ai/guides/code-intelligence.md): The code index, language servers, and architecture diagrams the model uses to understand a codebase, plus Workspace Radar and Project Scout.
- [Choosing a model](https://docs.open4rena.ai/models/overview.md): Pick and switch models, and keep the model list up to date.
- [Providers and API keys](https://docs.open4rena.ai/models/providers.md): The model providers o4 supports, and how to add API keys.
- [Claude and ChatGPT subscriptions](https://docs.open4rena.ai/models/subscriptions.md): Use a Claude Code or ChatGPT/Codex subscription instead of an API key.
- [Local models](https://docs.open4rena.ai/models/local-models.md): Run o4 against models on your own machine.
- [OpenAI-compatible endpoints](https://docs.open4rena.ai/models/openai-compatible.md): Connect o4 to any endpoint that speaks the OpenAI API.
- [Model router and fallback](https://docs.open4rena.ai/models/router.md): Switch to another model automatically when a provider is rate limited, overloaded, or failing.
- [Reasoning and prompt profiles](https://docs.open4rena.ai/models/reasoning.md): Control how hard a model thinks, and which prompt profile o4 uses.
- [Permissions](https://docs.open4rena.ai/safety/permissions.md): Choose what o4 may do without asking, with permission modes and rules.
- [Sandbox](https://docs.open4rena.ai/safety/sandbox.md): Limit what shell commands from the model can read, write, and reach.
- [Workspace trust](https://docs.open4rena.ai/safety/workspace-trust.md): Why o4 ignores a project's hooks and MCP servers until you trust it.
- [Configuration files](https://docs.open4rena.ai/configuration/overview.md): Where o4 keeps its settings, and which file wins.
- [Project instructions](https://docs.open4rena.ai/configuration/project-instructions.md): Give o4 standing instructions for a project with AGENTS.md and related files.
- [Themes and display](https://docs.open4rena.ai/configuration/appearance.md): Change o4's theme, key bindings, and display options.
- [MCP servers](https://docs.open4rena.ai/extend/mcp.md): Connect external tools to o4 with the Model Context Protocol.
- [Plugins and marketplaces](https://docs.open4rena.ai/extend/plugins.md): Install plugins that add commands, skills, agents, and more.
- [Skills](https://docs.open4rena.ai/extend/skills.md): Package reusable instructions that you or the model can load when they apply.
- [Hooks](https://docs.open4rena.ai/extend/hooks.md): Run your own commands when o4 reaches certain points in a session.
- [Harness guides and sensors](https://docs.open4rena.ai/extend/harness.md): Give o4 project-specific guides and check commands with .o4/harness.toml.
- [CodeMode (experimental)](https://docs.open4rena.ai/extend/codemode.md): Let the model write JavaScript programs that call o4's tools.
- [Language servers and formatters](https://docs.open4rena.ai/extend/lsp-and-formatters.md): Give the model code navigation through language servers, and a tool to format files.
- [The o4 daemon](https://docs.open4rena.ai/extend/daemon.md): Start, stop, and check the experimental o4 background daemon.
- [Evaluations](https://docs.open4rena.ai/extend/evals.md): Run o4 on a set of coding tasks and compare how models do.
- [Troubleshooting](https://docs.open4rena.ai/help/troubleshooting.md): Fix common problems with installing and running o4.
- [Telemetry and privacy](https://docs.open4rena.ai/help/telemetry-and-privacy.md): What o4 sends, what it never sends, and how to turn telemetry off.
- [Feedback and support](https://docs.open4rena.ai/help/feedback.md): Report a bug or send feedback about o4.
- [CLI reference](https://docs.open4rena.ai/reference/cli.md): Every o4 command-line command and option.
- [Slash commands](https://docs.open4rena.ai/reference/slash-commands.md): Every slash command you can type in o4.
- [Tools](https://docs.open4rena.ai/reference/tools.md): The tools the model can call while it works.
- [Keyboard shortcuts](https://docs.open4rena.ai/reference/keyboard-shortcuts.md): Keys for the prompt, the transcript, and dialogs.
- [Configuration reference](https://docs.open4rena.ai/reference/configuration.md): Every setting in o4's configuration files.
- [Environment variables](https://docs.open4rena.ai/reference/environment-variables.md): Environment variables that change how o4 behaves.