OpenAI - Codex Docs Changes 2026-04-23 00:46 UTC → 2026-04-24 12:28 UTC

103 103 

104For a middle ground, `approval_policy = { granular = { ... } }` lets you keep specific approval prompt categories interactive while automatically rejecting others. The granular policy covers sandbox approvals, execpolicy-rule prompts, MCP prompts, `request_permissions` prompts, and skill-script approvals.104For a middle ground, `approval_policy = { granular = { ... } }` lets you keep specific approval prompt categories interactive while automatically rejecting others. The granular policy covers sandbox approvals, execpolicy-rule prompts, MCP prompts, `request_permissions` prompts, and skill-script approvals.

105 105 

106Set `approvals_reviewer = "guardian_subagent"` to route eligible approval reviews through the Guardian reviewer subagent instead of prompting the user directly. Admin requirements can constrain this with `allowed_approvals_reviewers`.106### Automatic approval reviews

107 

108By default, approval requests route to you:

109 

110```toml

111approvals_reviewer = "user"

112```

113 

114Automatic approval reviews apply when approvals are interactive, such as

115`approval_policy = "on-request"` or a granular approval policy. Set

116`approvals_reviewer = "auto_review"` to route eligible approval requests

117through a reviewer agent before Codex runs the request:

118 

119```toml

120approval_policy = "on-request"

121approvals_reviewer = "auto_review"

122```

123 

124The reviewer evaluates only actions that already need approval, such as sandbox

125escalations, network requests, `request_permissions` prompts, or side-effecting

126app and MCP tool calls. Actions that stay inside the sandbox continue without an

127extra review step.

128 

129The reviewer policy checks for data exfiltration, credential probing, persistent

130security weakening, and destructive actions. Low-risk and medium-risk actions

131can proceed when policy allows them. The policy denies critical-risk actions.

132High-risk actions require enough user authorization and no matching deny rule.

133Timeouts, parse failures, and review errors fail closed.

134 

135The [default reviewer policy](https://github.com/openai/codex/blob/main/codex-rs/core/src/guardian/policy.md)

136is in the open-source Codex repository. Enterprises can replace its

137tenant-specific section with `guardian_policy_config` in managed requirements.

138Local `[auto_review].policy` text is also supported, but managed requirements

139take precedence. For setup details, see

140[Managed configuration](https://developers.openai.com/codex/enterprise/managed-configuration#configure-automatic-review-policy).

141 

142In the Codex app, these reviews appear as automatic review items with a status such

143as Reviewing, Approved, Denied, Stopped, or Timed out. They can also include a

144risk level for the reviewed request.

145 

146Automatic review uses extra model calls, so it can add to Codex usage. Admins

147can constrain it with `allowed_approvals_reviewers`.

107 148 

108### Common sandbox and approval combinations149### Common sandbox and approval combinations

109 150 

66 66 

67Run commands in each thread and launch repeatable project actions.](https://developers.openai.com/codex/app/features#integrated-terminal)[### In-app browser67Run commands in each thread and launch repeatable project actions.](https://developers.openai.com/codex/app/features#integrated-terminal)[### In-app browser

68 68 

69Open unauthenticated local or public pages and comment on rendered output.](https://developers.openai.com/codex/app/browser)[### Image generation69Open rendered pages, leave comments, or let Codex operate local browser flows.](https://developers.openai.com/codex/app/browser)[### Image generation

70 70 

71Generate or edit images in a thread while you work on the surrounding code and assets.](https://developers.openai.com/codex/app/features#image-generation)[### Automations71Generate or edit images in a thread while you work on the surrounding code and assets.](https://developers.openai.com/codex/app/features#image-generation)[### Automations

72 72 

20 20 

21![Codex app showing a browser comment on a local web app preview](/images/codex/app/in-app-browser-light.webp)21![Codex app showing a browser comment on a local web app preview](/images/codex/app/in-app-browser-light.webp)

22 22 

23## Browser use

24 

25Browser use lets Codex operate the in-app browser directly. Use it for local

26development servers and file-backed previews when Codex needs to click, type,

27inspect rendered state, take screenshots, or verify a fix in the page.

28 

29To use it, install and enable the Browser plugin. Then ask Codex to use the

30browser in your task, or reference it directly with `@Browser`. The app keeps

31browser use inside the in-app browser and lets you manage allowed and blocked

32websites from settings.

33 

34Example:

35 

36```text

37Use the browser to open http://localhost:3000/settings, reproduce the layout

38bug, and fix only the overflowing controls.

39```

40 

41Codex asks before using a website unless you've allowed it. Removing a site from

42the allowed list means Codex asks again before using it; removing a site from the

43blocked list means Codex can ask again instead of treating it as blocked.

44 

23## Preview a page45## Preview a page

24 46 

251. Start your app's development server in the [integrated terminal](https://developers.openai.com/codex/app/features#integrated-terminal) or with a [local environment action](https://developers.openai.com/codex/app/local-environments#actions).471. Start your app's development server in the [integrated terminal](https://developers.openai.com/codex/app/features#integrated-terminal) or with a [local environment action](https://developers.openai.com/codex/app/local-environments#actions).

4with built-in worktree support, automations, and Git functionality.4with built-in worktree support, automations, and Git functionality.

5 5 

6Most Codex app features are available on both macOS and Windows.6Most Codex app features are available on both macOS and Windows.

7Platform-specific exceptions are noted below.7The sections below note platform-specific exceptions.

8 8 

9---9---

10 10 

146Use browser comments to mark specific elements or areas on a page, then ask146Use browser comments to mark specific elements or areas on a page, then ask

147Codex to address that feedback.147Codex to address that feedback.

148 148 

149When you want Codex to operate the page directly, use

150[browser use](https://developers.openai.com/codex/app/browser#browser-use) for local development servers and

151file-backed pages. You can manage the Browser plugin, allowed websites, and

152blocked websites from settings.

153 

149![Codex app showing a browser comment on a local web app preview](/images/codex/app/in-app-browser-light.webp)154![Codex app showing a browser comment on a local web app preview](/images/codex/app/in-app-browser-light.webp)

150 155 

151## Computer use156## Computer use

225opening separate projects or using worktrees rather than asking Codex to roam230opening separate projects or using worktrees rather than asking Codex to roam

226outside the project root.231outside the project root.

227 232 

233If [automatic review](https://developers.openai.com/codex/agent-approvals-security#automatic-approval-reviews)

234is available in your workspace, you can choose it from the permissions selector.

235It keeps the same sandbox boundary but routes eligible approval requests through

236the configured review policy instead of waiting for you.

237 

228For a high-level overview, see [sandboxing](https://developers.openai.com/codex/concepts/sandboxing). For238For a high-level overview, see [sandboxing](https://developers.openai.com/codex/concepts/sandboxing). For

229configuration details, see the239configuration details, see the

230[agent approvals & security documentation](https://developers.openai.com/codex/agent-approvals-security).240[agent approvals & security documentation](https://developers.openai.com/codex/agent-approvals-security).

43also apply to the Codex CLI and IDE extension because the MCP configuration lives in43also apply to the Codex CLI and IDE extension because the MCP configuration lives in

44`config.toml`. See the [Model Context Protocol docs](https://developers.openai.com/codex/mcp) for details.44`config.toml`. See the [Model Context Protocol docs](https://developers.openai.com/codex/mcp) for details.

45 45 

46## Browser use

47 

48Use these settings to install or enable the bundled Browser plugin and manage

49allowlisted and blocklisted websites. Codex asks before using a website

50unless you’ve allowlisted it. Removing a site from the blocklist lets Codex ask

51again before using it in the browser.

52 

53See [In-app browser](https://developers.openai.com/codex/app/browser) for browser preview, comment, and

54browser use workflows.

55 

46## Computer Use56## Computer Use

47 57 

48On macOS, check your Computer Use settings to review desktop-app access and related58On macOS, check your Computer Use settings to review desktop-app access and related

43 43 

44 npm i -g @openai/codex@latestCopy44 npm i -g @openai/codex@latestCopy

45 45 

46The Codex CLI is available on macOS and Linux. Windows support is46The Codex CLI is available on macOS, Windows, and Linux. On Windows, run Codex

47experimental. For the best Windows experience, use Codex in a WSL2 workspace47 natively in PowerShell with the Windows sandbox, or use WSL2 when you need a

48and follow our [Windows setup guide](https://developers.openai.com/codex/windows).48Linux-native environment. For setup details, see the

49[Windows setup guide](https://developers.openai.com/codex/windows).

49 50 

50If you're new to Codex, read the [best practices guide](https://developers.openai.com/codex/learn/best-practices).51If you're new to Codex, read the [best practices guide](https://developers.openai.com/codex/learn/best-practices).

51 52 

107 107 

108## Models and reasoning108## Models and reasoning

109 109 

110For most tasks in Codex, `gpt-5.4` is the recommended model. It brings the110For most tasks in Codex, `gpt-5.5` is the recommended model when it is

111industry-leading coding capabilities of `gpt-5.3-codex` to OpenAI’s flagship111available. It is OpenAI's newest frontier model for complex coding, computer

112frontier model, combining frontier coding performance with stronger reasoning,112use, knowledge work, and research workflows, with stronger planning, tool use,

113native computer use, and broader professional workflows. For extra fast tasks,113and follow-through on multi-step tasks. If `gpt-5.5` is not yet available,

114ChatGPT Pro subscribers have access to the GPT-5.3-Codex-Spark model in114continue using `gpt-5.4`. For extra fast tasks, ChatGPT Pro subscribers have

115research preview.115access to the GPT-5.3-Codex-Spark model in research preview.

116 116 

117Switch models mid-session with the `/model` command, or specify one when launching the CLI.117Switch models mid-session with the `/model` command, or specify one when launching the CLI.

118 118 

119```bash119```bash

120codex --model gpt-5.4120codex --model gpt-5.5

121```121```

122 122 

123[Learn more about the models available in Codex](https://developers.openai.com/codex/models).123[Learn more about the models available in Codex](https://developers.openai.com/codex/models).

40| [`/mcp`](#list-mcp-tools-with-mcp) | List configured Model Context Protocol (MCP) tools. | Check which external tools Codex can call during the session. |40| [`/mcp`](#list-mcp-tools-with-mcp) | List configured Model Context Protocol (MCP) tools. | Check which external tools Codex can call during the session. |

41| [`/mention`](#highlight-files-with-mention) | Attach a file to the conversation. | Point Codex at specific files or folders you want it to inspect next. |41| [`/mention`](#highlight-files-with-mention) | Attach a file to the conversation. | Point Codex at specific files or folders you want it to inspect next. |

42| [`/model`](#set-the-active-model-with-model) | Choose the active model (and reasoning effort, when available). | Switch between general-purpose models (`gpt-4.1-mini`) and deeper reasoning models before running a task. |42| [`/model`](#set-the-active-model-with-model) | Choose the active model (and reasoning effort, when available). | Switch between general-purpose models (`gpt-4.1-mini`) and deeper reasoning models before running a task. |

43| [`/fast`](#toggle-fast-mode-with-fast) | Toggle Fast mode for GPT-5.4. | Turn Fast mode on or off, or check whether the current thread is using it. |43| [`/fast`](#toggle-fast-mode-with-fast) | Toggle Fast mode for supported models. | Turn Fast mode on or off, or check whether the current thread is using it. |

44| [`/plan`](#switch-to-plan-mode-with-plan) | Switch to plan mode and optionally send a prompt. | Ask Codex to propose an execution plan before implementation work starts. |44| [`/plan`](#switch-to-plan-mode-with-plan) | Switch to plan mode and optionally send a prompt. | Ask Codex to propose an execution plan before implementation work starts. |

45| [`/personality`](#set-a-communication-style-with-personality) | Choose a communication style for responses. | Make Codex more concise, more explanatory, or more collaborative without changing your instructions. |45| [`/personality`](#set-a-communication-style-with-personality) | Choose a communication style for responses. | Make Codex more concise, more explanatory, or more collaborative without changing your instructions. |

46| [`/ps`](#check-background-terminals-with-ps) | Show experimental background terminals and their recent output. | Check long-running commands without leaving the main transcript. |46| [`/ps`](#check-background-terminals-with-ps) | Show experimental background terminals and their recent output. | Check long-running commands without leaving the main transcript. |

170[Codex app features](https://developers.openai.com/codex/app/features#approvals-and-sandboxing), and for the170[Codex app features](https://developers.openai.com/codex/app/features#approvals-and-sandboxing), and for the

171IDE-specific settings entry points, see [Codex IDE extension settings](https://developers.openai.com/codex/ide/settings).171IDE-specific settings entry points, see [Codex IDE extension settings](https://developers.openai.com/codex/ide/settings).

172 172 

173Automatic review, when available, doesn't change the sandbox boundary. It

174reviews approval requests, such as sandbox escalations or network access, while

175actions already allowed inside the sandbox run without extra review. See

176[Automatic approval reviews](https://developers.openai.com/codex/agent-approvals-security#automatic-approval-reviews)

177for the policy behavior.

178 

173Platform details live in the platform-specific docs. For native Windows setup,179Platform details live in the platform-specific docs. For native Windows setup,

174behavior, and troubleshooting, see [Windows](https://developers.openai.com/codex/windows). For admin180behavior, and troubleshooting, see [Windows](https://developers.openai.com/codex/windows). For admin

175requirements and organization-level constraints on sandboxing and approvals, see181requirements and organization-level constraints on sandboxing and approvals, see

65 65 

66If you don't pin a model or `model_reasoning_effort`, Codex can choose a setup66If you don't pin a model or `model_reasoning_effort`, Codex can choose a setup

67that balances intelligence, speed, and price for the task. It may favor67that balances intelligence, speed, and price for the task. It may favor

68`gpt-5.4-mini` for fast scans or a higher-effort `gpt-5.4`68`gpt-5.4-mini` for fast scans or a higher-effort `gpt-5.5` configuration for

69configuration for more demanding reasoning. When you want finer control, steer that69more demanding reasoning when that model is available. When you want finer

70choice in your prompt or set `model` and `model_reasoning_effort` directly in70control, steer that choice in your prompt or set `model` and

71the agent file.71`model_reasoning_effort` directly in the agent file.

72 72 

73For most tasks in Codex, start with `gpt-5.4`. Use `gpt-5.4-mini` when you73For most tasks in Codex, start with `gpt-5.5` when it is available. Continue

74want a faster, lower-cost option for lighter subagent work. If you have74 using `gpt-5.4` during the rollout if `gpt-5.5` is not yet available. Use

75ChatGPT Pro and want near-instant text-only iteration, `gpt-5.3-codex-spark`75 `gpt-5.4-mini` when you want a faster, lower-cost option for lighter subagent

76remains available in research preview.76 work. If you have ChatGPT Pro and want near-instant text-only iteration,

77 `gpt-5.3-codex-spark` remains available in research preview.

77 78 

78### Model choice79### Model choice

79 80 

80- **`gpt-5.4`**: Start here for most agents. It combines strong coding, reasoning, tool use, and broader workflows. The main agent and agents that coordinate ambiguous or multi-step work fit here.81- **`gpt-5.5`**: Start here for demanding agents when it is available. It is strongest for ambiguous, multi-step work that needs planning, tool use, validation, and follow-through across a larger context.

82- **`gpt-5.4`**: Use this when `gpt-5.5` is not yet available or when a workflow is pinned to GPT-5.4. It combines strong coding, reasoning, tool use, and broader workflows.

81- **`gpt-5.4-mini`**: Use for agents that favor speed and efficiency over depth, such as exploration, read-heavy scans, large-file review, or processing supporting documents. It works well for parallel workers that return distilled results to the main agent.83- **`gpt-5.4-mini`**: Use for agents that favor speed and efficiency over depth, such as exploration, read-heavy scans, large-file review, or processing supporting documents. It works well for parallel workers that return distilled results to the main agent.

82- **`gpt-5.3-codex-spark`**: If you have ChatGPT Pro, use this research preview model for near-instant, text-only iteration when latency matters more than broader capability.84- **`gpt-5.3-codex-spark`**: If you have ChatGPT Pro, use this research preview model for near-instant, text-only iteration when latency matters more than broader capability.

83 85 

230 230 

231You can also use a granular approval policy (`approval_policy = { granular = { ... } }`) to allow or auto-reject individual prompt categories. This is useful when you want normal interactive approvals for some cases but want others, such as `request_permissions` or skill-script prompts, to fail closed automatically.231You can also use a granular approval policy (`approval_policy = { granular = { ... } }`) to allow or auto-reject individual prompt categories. This is useful when you want normal interactive approvals for some cases but want others, such as `request_permissions` or skill-script prompts, to fail closed automatically.

232 232 

233```233Set `approvals_reviewer = "auto_review"` to route eligible interactive approval

234requests through automatic review. This changes the reviewer, not the sandbox

235boundary.

236 

237Use `[auto_review].policy` for local reviewer policy instructions. Managed

238`guardian_policy_config` takes precedence.

239 

240```toml

234approval_policy = "untrusted" # Other options: on-request, never, or { granular = { ... } }241approval_policy = "untrusted" # Other options: on-request, never, or { granular = { ... } }

242approvals_reviewer = "user" # Or "auto_review" for automatic review

235sandbox_mode = "workspace-write"243sandbox_mode = "workspace-write"

236allow_login_shell = false # Optional hardening: disallow login shells for shell tools244allow_login_shell = false # Optional hardening: disallow login shells for shell tools

237 245 

249exclude_slash_tmp = false # Allow /tmp257exclude_slash_tmp = false # Allow /tmp

250writable_roots = ["/Users/YOU/.pyenv/shims"]258writable_roots = ["/Users/YOU/.pyenv/shims"]

251network_access = false # Opt in to outbound network259network_access = false # Opt in to outbound network

260 

261[auto_review]

262policy = """

263Use your organization's automatic review policy.

264"""

252```265```

253 266 

254Need the complete key list (including profile-scoped overrides and requirements constraints)? See [Configuration Reference](https://developers.openai.com/codex/config-reference) and [Managed configuration](https://developers.openai.com/codex/enterprise/managed-configuration).267Need the complete key list (including profile-scoped overrides and requirements constraints)? See [Configuration Reference](https://developers.openai.com/codex/config-reference) and [Managed configuration](https://developers.openai.com/codex/enterprise/managed-configuration).

46Choose the model Codex uses by default in the CLI and IDE.46Choose the model Codex uses by default in the CLI and IDE.

47 47 

48```toml48```toml

49model = "gpt-5.4"49model = "gpt-5.5"

50```50```

51 51 

52#### Approval prompts52#### Approval prompts

155| `personality` | true | Stable | Enable personality selection controls |155| `personality` | true | Stable | Enable personality selection controls |

156| `shell_snapshot` | true | Stable | Snapshot your shell environment to speed up repeated commands |156| `shell_snapshot` | true | Stable | Snapshot your shell environment to speed up repeated commands |

157| `shell_tool` | true | Stable | Enable the default `shell` tool |157| `shell_tool` | true | Stable | Enable the default `shell` tool |

158| `guardian_approval` | false | Experimental | Route eligible approval requests through the guardian reviewer subagent (set `approvals_reviewer = "guardian_subagent"`). |

159| `unified_exec` | `true` except Windows | Stable | Use the unified PTY-backed exec tool |158| `unified_exec` | `true` except Windows | Stable | Use the unified PTY-backed exec tool |

160| `undo` | false | Stable | Enable undo via per-turn git ghost snapshots |159| `undo` | false | Stable | Enable undo via per-turn git ghost snapshots |

161| `web_search` | true | Deprecated | Legacy toggle; prefer the top-level `web_search` setting |160| `web_search` | true | Deprecated | Legacy toggle; prefer the top-level `web_search` setting |

24| `approval_policy.granular.rules` | `boolean` | When `true`, approvals triggered by execpolicy `prompt` rules are allowed to surface. |24| `approval_policy.granular.rules` | `boolean` | When `true`, approvals triggered by execpolicy `prompt` rules are allowed to surface. |

25| `approval_policy.granular.sandbox_approval` | `boolean` | When `true`, sandbox escalation approval prompts are allowed to surface. |25| `approval_policy.granular.sandbox_approval` | `boolean` | When `true`, sandbox escalation approval prompts are allowed to surface. |

26| `approval_policy.granular.skill_approval` | `boolean` | When `true`, skill-script approval prompts are allowed to surface. |26| `approval_policy.granular.skill_approval` | `boolean` | When `true`, skill-script approval prompts are allowed to surface. |

27| `approvals_reviewer` | `user | guardian_subagent` | Select who reviews eligible approval prompts. Defaults to `user`; `guardian_subagent` routes supported reviews through the Guardian reviewer subagent. |27| `approvals_reviewer` | `user | auto_review` | Who reviews eligible approval prompts under `on-request` or granular approval policies. Defaults to `user`; `auto_review` uses the reviewer subagent. This setting doesn't change sandboxing or review actions already allowed inside the sandbox. |

28| `apps._default.destructive_enabled` | `boolean` | Default allow/deny for app tools with `destructive_hint = true`. |28| `apps._default.destructive_enabled` | `boolean` | Default allow/deny for app tools with `destructive_hint = true`. |

29| `apps._default.enabled` | `boolean` | Default app enabled state for all apps unless overridden per app. |29| `apps._default.enabled` | `boolean` | Default app enabled state for all apps unless overridden per app. |

30| `apps._default.open_world_enabled` | `boolean` | Default allow/deny for app tools with `open_world_hint = true`. |30| `apps._default.open_world_enabled` | `boolean` | Default allow/deny for app tools with `open_world_hint = true`. |

35| `apps.<id>.open_world_enabled` | `boolean` | Allow or block tools in this app that advertise `open_world_hint = true`. |35| `apps.<id>.open_world_enabled` | `boolean` | Allow or block tools in this app that advertise `open_world_hint = true`. |

36| `apps.<id>.tools.<tool>.approval_mode` | `auto | prompt | approve` | Per-tool approval behavior override for a single app tool. |36| `apps.<id>.tools.<tool>.approval_mode` | `auto | prompt | approve` | Per-tool approval behavior override for a single app tool. |

37| `apps.<id>.tools.<tool>.enabled` | `boolean` | Per-tool enabled override for an app tool (for example `repos/list`). |37| `apps.<id>.tools.<tool>.enabled` | `boolean` | Per-tool enabled override for an app tool (for example `repos/list`). |

38| `auto_review.policy` | `string` | Local Markdown policy instructions for automatic review. Managed `guardian_policy_config` takes precedence. Blank values are ignored. |

38| `background_terminal_max_timeout` | `number` | Maximum poll window in milliseconds for empty `write_stdin` polls (background terminal polling). Default: `300000` (5 minutes). Replaces the older `background_terminal_timeout` key. |39| `background_terminal_max_timeout` | `number` | Maximum poll window in milliseconds for empty `write_stdin` polls (background terminal polling). Default: `300000` (5 minutes). Replaces the older `background_terminal_timeout` key. |

39| `chatgpt_base_url` | `string` | Override the base URL used during the ChatGPT login flow. |40| `chatgpt_base_url` | `string` | Override the base URL used during the ChatGPT login flow. |

40| `check_for_update_on_startup` | `boolean` | Check for Codex updates on startup (set to false only when updates are centrally managed). |41| `check_for_update_on_startup` | `boolean` | Check for Codex updates on startup (set to false only when updates are centrally managed). |

50| `features.codex_hooks` | `boolean` | Enable lifecycle hooks loaded from `hooks.json` (under development; off by default). |51| `features.codex_hooks` | `boolean` | Enable lifecycle hooks loaded from `hooks.json` (under development; off by default). |

51| `features.enable_request_compression` | `boolean` | Compress streaming request bodies with zstd when supported (stable; on by default). |52| `features.enable_request_compression` | `boolean` | Compress streaming request bodies with zstd when supported (stable; on by default). |

52| `features.fast_mode` | `boolean` | Enable Fast mode selection and the `service_tier = "fast"` path (stable; on by default). |53| `features.fast_mode` | `boolean` | Enable Fast mode selection and the `service_tier = "fast"` path (stable; on by default). |

53| `features.guardian_approval` | `boolean` | Route eligible approval requests through the guardian reviewer subagent (experimental; off by default). Use with `approvals_reviewer = "guardian_subagent"`. |

54| `features.memories` | `boolean` | Enable [Memories](https://developers.openai.com/codex/memories) (off by default). |54| `features.memories` | `boolean` | Enable [Memories](https://developers.openai.com/codex/memories) (off by default). |

55| `features.multi_agent` | `boolean` | Enable multi-agent collaboration tools (`spawn_agent`, `send_input`, `resume_agent`, `wait_agent`, and `close_agent`) (stable; on by default). |55| `features.multi_agent` | `boolean` | Enable multi-agent collaboration tools (`spawn_agent`, `send_input`, `resume_agent`, `wait_agent`, and `close_agent`) (stable; on by default). |

56| `features.personality` | `boolean` | Enable personality selection controls (stable; on by default). |56| `features.personality` | `boolean` | Enable personality selection controls (stable; on by default). |

104| `memories.max_unused_days` | `number` | Maximum days since a memory was last used before it becomes ineligible for consolidation. Defaults to `30` and is clamped to `0`-`365`. |104| `memories.max_unused_days` | `number` | Maximum days since a memory was last used before it becomes ineligible for consolidation. Defaults to `30` and is clamped to `0`-`365`. |

105| `memories.min_rollout_idle_hours` | `number` | Minimum idle time before a thread is considered for memory generation. Defaults to `6` and is clamped to `1`-`48`. |105| `memories.min_rollout_idle_hours` | `number` | Minimum idle time before a thread is considered for memory generation. Defaults to `6` and is clamped to `1`-`48`. |

106| `memories.use_memories` | `boolean` | When `false`, Codex skips injecting existing memories into future sessions. Defaults to `true`. |106| `memories.use_memories` | `boolean` | When `false`, Codex skips injecting existing memories into future sessions. Defaults to `true`. |

107| `model` | `string` | Model to use (e.g., `gpt-5.4`). |107| `model` | `string` | Model to use (e.g., `gpt-5.5`). |

108| `model_auto_compact_token_limit` | `number` | Token threshold that triggers automatic history compaction (unset uses model defaults). |108| `model_auto_compact_token_limit` | `number` | Token threshold that triggers automatic history compaction (unset uses model defaults). |

109| `model_catalog_json` | `string (path)` | Optional path to a JSON model catalog loaded on startup. Profile-level `profiles.<name>.model_catalog_json` can override this per profile. |109| `model_catalog_json` | `string (path)` | Optional path to a JSON model catalog loaded on startup. Profile-level `profiles.<name>.model_catalog_json` can override this per profile. |

110| `model_context_window` | `number` | Context window tokens available to the active model. |110| `model_context_window` | `number` | Context window tokens available to the active model. |

409 409 

410Type / Values410Type / Values

411 411 

412`user | guardian_subagent`412`user | auto_review`

413 413 

414Details414Details

415 415 

416Select who reviews eligible approval prompts. Defaults to `user`; `guardian_subagent` routes supported reviews through the Guardian reviewer subagent.416Who reviews eligible approval prompts under `on-request` or granular approval policies. Defaults to `user`; `auto_review` uses the reviewer subagent. This setting doesn't change sandboxing or review actions already allowed inside the sandbox.

417 417 

418Key418Key

419 419 

537 537 

538Key538Key

539 539 

540`auto_review.policy`

541 

542Type / Values

543 

544`string`

545 

546Details

547 

548Local Markdown policy instructions for automatic review. Managed `guardian_policy_config` takes precedence. Blank values are ignored.

549 

550Key

551 

540`background_terminal_max_timeout`552`background_terminal_max_timeout`

541 553 

542Type / Values554Type / Values

717 729 

718Key730Key

719 731 

720`features.guardian_approval`

721 

722Type / Values

723 

724`boolean`

725 

726Details

727 

728Route eligible approval requests through the guardian reviewer subagent (experimental; off by default). Use with `approvals_reviewer = "guardian_subagent"`.

729 

730Key

731 

732`features.memories`732`features.memories`

733 733 

734Type / Values734Type / Values

1373 1373 

1374Details1374Details

1375 1375 

1376Model to use (e.g., `gpt-5.4`).1376Model to use (e.g., `gpt-5.5`).

1377 1377 

1378Key1378Key

1379 1379 

2948| Key | Type / Values | Details |2948| Key | Type / Values | Details |

2949| --- | --- | --- |2949| --- | --- | --- |

2950| `allowed_approval_policies` | `array<string>` | Allowed values for `approval_policy` (for example `untrusted`, `on-request`, `never`, and `granular`). |2950| `allowed_approval_policies` | `array<string>` | Allowed values for `approval_policy` (for example `untrusted`, `on-request`, `never`, and `granular`). |

2951| `allowed_approvals_reviewers` | `array<string>` | Allowed values for `approvals_reviewer` (for example `user` and `guardian_subagent`). |2951| `allowed_approvals_reviewers` | `array<string>` | Allowed values for `approvals_reviewer`, such as `user` and `auto_review`. |

2952| `allowed_sandbox_modes` | `array<string>` | Allowed values for `sandbox_mode`. |2952| `allowed_sandbox_modes` | `array<string>` | Allowed values for `sandbox_mode`. |

2953| `allowed_web_search_modes` | `array<string>` | Allowed values for `web_search` (`disabled`, `cached`, `live`). `disabled` is always allowed; an empty list effectively allows only `disabled`. |2953| `allowed_web_search_modes` | `array<string>` | Allowed values for `web_search` (`disabled`, `cached`, `live`). `disabled` is always allowed; an empty list effectively allows only `disabled`. |

2954| `features` | `table` | Pinned feature values keyed by the canonical names from `config.toml`'s `[features]` table. |2954| `features` | `table` | Pinned feature values keyed by the canonical names from `config.toml`'s `[features]` table. |

2955| `features.<name>` | `boolean` | Require a specific canonical feature key to stay enabled or disabled. |2955| `features.<name>` | `boolean` | Require a specific canonical feature key to stay enabled or disabled. |

2956| `guardian_policy_config` | `string` | Managed Markdown policy instructions for automatic review. This takes precedence over local `[auto_review].policy`. Blank values are ignored. |

2956| `mcp_servers` | `table` | Allowlist of MCP servers that may be enabled. Both the server name (`<id>`) and its identity must match for the MCP server to be enabled. Any configured MCP server not in the allowlist (or with a mismatched identity) is disabled. |2957| `mcp_servers` | `table` | Allowlist of MCP servers that may be enabled. Both the server name (`<id>`) and its identity must match for the MCP server to be enabled. Any configured MCP server not in the allowlist (or with a mismatched identity) is disabled. |

2957| `mcp_servers.<id>.identity` | `table` | Identity rule for a single MCP server. Set either `command` (stdio) or `url` (streamable HTTP). |2958| `mcp_servers.<id>.identity` | `table` | Identity rule for a single MCP server. Set either `command` (stdio) or `url` (streamable HTTP). |

2958| `mcp_servers.<id>.identity.command` | `string` | Allow an MCP stdio server when its `mcp_servers.<id>.command` matches this command. |2959| `mcp_servers.<id>.identity.command` | `string` | Allow an MCP stdio server when its `mcp_servers.<id>.command` matches this command. |

2988 2989 

2989Details2990Details

2990 2991 

2991Allowed values for `approvals_reviewer` (for example `user` and `guardian_subagent`).2992Allowed values for `approvals_reviewer`, such as `user` and `auto_review`.

2992 2993 

2993Key2994Key

2994 2995 

3040 3041 

3041Key3042Key

3042 3043 

3044`guardian_policy_config`

3045 

3046Type / Values

3047 

3048`string`

3049 

3050Details

3051 

3052Managed Markdown policy instructions for automatic review. This takes precedence over local `[auto_review].policy`. Blank values are ignored.

3053 

3054Key

3055 

3043`mcp_servers`3056`mcp_servers`

3044 3057 

3045Type / Values3058Type / Values

27# Core Model Selection27# Core Model Selection

28################################################################################28################################################################################

29 29 

30# Primary model used by Codex. Recommended example for most users: "gpt-5.4".30# Primary model used by Codex. Recommended example for most users: "gpt-5.5".

31model = "gpt-5.4"31model = "gpt-5.5"

32 32 

33# Communication style for supported models. Allowed values: none | friendly | pragmatic33# Communication style for supported models. Allowed values: none | friendly | pragmatic

34# personality = "pragmatic"34# personality = "pragmatic"

35 35 

36# Optional model override for /review. Default: unset (uses current session model).36# Optional model override for /review. Default: unset (uses current session model).

37# review_model = "gpt-5.4"37# review_model = "gpt-5.5"

38 38 

39# Provider id selected from [model_providers]. Default: "openai".39# Provider id selected from [model_providers]. Default: "openai".

40model_provider = "openai"40model_provider = "openai"

109# - never: never prompt (risky)109# - never: never prompt (risky)

110# - { granular = { ... } }: allow or auto-reject selected prompt categories110# - { granular = { ... } }: allow or auto-reject selected prompt categories

111approval_policy = "on-request"111approval_policy = "on-request"

112# Who reviews eligible approval prompts: user (default) | guardian_subagent112# Who reviews eligible approval prompts: user (default) | auto_review

113# approvals_reviewer = "user"113# approvals_reviewer = "user"

114 114 

115# Example granular policy:115# Example granular policy:

393# multi_agent = true393# multi_agent = true

394# personality = true394# personality = true

395# fast_mode = true395# fast_mode = true

396# guardian_approval = false

397# enable_request_compression = true396# enable_request_compression = true

398# skill_mcp_dependency_install = true397# skill_mcp_dependency_install = true

399# prevent_idle_sleep = false398# prevent_idle_sleep = false

7 7 

8## Admin-enforced requirements (requirements.toml)8## Admin-enforced requirements (requirements.toml)

9 9 

10Requirements constrain security-sensitive settings (approval policy, sandbox mode, web search mode, and optionally which MCP servers users can enable). When resolving configuration (for example from `config.toml`, profiles, or CLI config overrides), if a value conflicts with an enforced rule, Codex falls back to a compatible value and notifies the user. If you configure an `mcp_servers` allowlist, Codex enables an MCP server only when both its name and identity match an approved entry; otherwise, Codex disables it.10Requirements constrain security-sensitive settings (approval policy, approvals reviewer, automatic review policy, sandbox mode, web search mode, and optionally which MCP servers users can enable). When resolving configuration (for example from `config.toml`, profiles, or CLI config overrides), if a value conflicts with an enforced rule, Codex falls back to a compatible value and notifies the user. If you configure an `mcp_servers` allowlist, Codex enables an MCP server only when both its name and identity match an approved entry; otherwise, Codex disables it.

11 11 

12Requirements can also constrain [feature flags](https://developers.openai.com/codex/config-basic/#feature-flags) via the `[features]` table in `requirements.toml`. Note that features aren't always security-sensitive, but enterprises can pin values if desired. Omitted keys remain unconstrained.12Requirements can also constrain [feature flags](https://developers.openai.com/codex/config-basic/#feature-flags) via the `[features]` table in `requirements.toml`. Note that features aren't always security-sensitive, but enterprises can pin values if desired. Omitted keys remain unconstrained.

13 13 

91 91 

92Use the canonical feature keys from `config.toml`'s `[features]` table. Codex normalizes the resulting feature set to meet these pins and rejects conflicting writes to `config.toml` or profile-scoped feature settings.92Use the canonical feature keys from `config.toml`'s `[features]` table. Codex normalizes the resulting feature set to meet these pins and rejects conflicting writes to `config.toml` or profile-scoped feature settings.

93 93 

94### Configure automatic review policy

95 

96Use `allowed_approvals_reviewers` to require or allow automatic review. Set it

97to `["auto_review"]` to require automatic review, or include `"user"` when users

98can choose manual approval.

99 

100Set `guardian_policy_config` to replace the tenant-specific section of the

101automatic review policy. Codex still uses the built-in reviewer template and

102output contract. Managed `guardian_policy_config` takes precedence over local

103`[auto_review].policy`.

104 

105```toml

106allowed_approval_policies = ["on-request"]

107allowed_approvals_reviewers = ["auto_review"]

108 

109guardian_policy_config = """

110## Environment Profile

111- Trusted internal destinations include github.com/my-org, artifacts.example.com,

112 and internal CI systems.

113 

114## Tenant Risk Taxonomy and Allow/Deny Rules

115- Treat uploads to unapproved third-party file-sharing services as high risk.

116- Deny actions that expose credentials or private source code to untrusted

117 destinations.

118"""

119```

120 

94### Enforce deny-read requirements121### Enforce deny-read requirements

95 122 

96Admins can deny reads for exact paths or glob patterns with123Admins can deny reads for exact paths or glob patterns with

106```133```

107 134 

108When deny-read requirements are present, Codex constrains local sandbox mode to135When deny-read requirements are present, Codex constrains local sandbox mode to

109`read-only` or `workspace-write` so the requirement can be enforced. On native136`read-only` or `workspace-write` so Codex can enforce them. On native

110Windows, managed `deny_read` applies to direct file tools; shell subprocess137Windows, managed `deny_read` applies to direct file tools; shell subprocess

111reads don’t use this sandbox requirement.138reads don't use this sandbox rule.

112 139 

113### Enforce command rules from requirements140### Enforce command rules from requirements

114 141 

16- [Download for Visual Studio Code Insiders](https://marketplace.visualstudio.com/items?itemName=openai.chatgpt)16- [Download for Visual Studio Code Insiders](https://marketplace.visualstudio.com/items?itemName=openai.chatgpt)

17- [Download for JetBrains IDEs](#jetbrains-ide-integration)17- [Download for JetBrains IDEs](#jetbrains-ide-integration)

18 18 

19The Codex VS Code extension is available on macOS and Linux. Windows support19Codex IDE integrations for VS Code-compatible editors and JetBrains IDEs are

20is experimental. For the best Windows experience, use Codex in a WSL220 available on macOS, Windows, and Linux. On Windows, run Codex natively with

21workspace and follow our [Windows setup guide](https://developers.openai.com/codex/windows).21 the Windows sandbox, or use WSL2 when you need a Linux-native environment. For

22setup details, see the [Windows setup guide](https://developers.openai.com/codex/windows).

22 23 

23After you install it, you'll find Codex in your editor sidebar.24After you install it, you'll find Codex in your editor sidebar.

24In VS Code, Codex opens in the right sidebar by default.25In VS Code, Codex opens in the right sidebar by default.

24| `chatgpt.commentCodeLensEnabled` | Show CodeLens above to-do comments so you can complete them with Codex. |24| `chatgpt.commentCodeLensEnabled` | Show CodeLens above to-do comments so you can complete them with Codex. |

25| `chatgpt.localeOverride` | Preferred language for the Codex UI. Leave empty to detect automatically. |25| `chatgpt.localeOverride` | Preferred language for the Codex UI. Leave empty to detect automatically. |

26| `chatgpt.openOnStartup` | Focus the Codex sidebar when the extension finishes starting. |26| `chatgpt.openOnStartup` | Focus the Codex sidebar when the extension finishes starting. |

27| `chatgpt.runCodexInWindowsSubsystemForLinux` | Windows only: Run Codex in WSL when Windows Subsystem for Linux (WSL) is available. Recommended for improved sandbox security and better performance. Codex agent mode on Windows currently requires WSL. Changing this setting reloads VS Code to apply the change. |27| `chatgpt.runCodexInWindowsSubsystemForLinux` | Windows only: Run Codex in WSL when Windows Subsystem for Linux (WSL) is available. Use this when your repositories and tooling live in WSL2 or when you need Linux-native tooling. Otherwise, Codex can run natively on Windows with the Windows sandbox. Changing this setting reloads VS Code to apply the change. |

2 2 

3## Recommended models3## Recommended models

4 4 

5![gpt-5.5](/images/api/models/gpt-5.5.jpg)

6 

7gpt-5.5

8 

9OpenAI's newest frontier model for complex coding, computer use, knowledge work, and research workflows in Codex.

10 

11codex -m gpt-5.5

12 

13Copy command

14 

15Capability

16 

17Speed

18 

19Codex CLI & SDK

20 

21Codex app & IDE extension

22 

23Codex Cloud

24 

25ChatGPT Credits

26 

27API Access

28 

5![gpt-5.4](/images/api/models/gpt-5.4.jpg)29![gpt-5.4](/images/api/models/gpt-5.4.jpg)

6 30 

7gpt-5.431gpt-5.4

98 122 

99API Access123API Access

100 124 

101For most tasks in Codex, start with `gpt-5.4`. It combines strong coding,125For most tasks in Codex, start with `gpt-5.5` when it appears in your model

102reasoning, native computer use, and broader professional workflows in one126 picker. It is strongest for complex coding, computer use, knowledge work, and

103model. Use `gpt-5.4-mini` when you want a faster, lower-cost option for127 research workflows. GPT-5.5 is currently available in Codex when you sign in

104lighter coding tasks or subagents. The `gpt-5.3-codex-spark` model is128 with ChatGPT; it isn't available with API-key authentication. During the

105available in research preview for ChatGPT Pro subscribers and is optimized for129 rollout, continue using `gpt-5.4` if `gpt-5.5` is not yet available. Use

106near-instant, real-time coding iteration.130 `gpt-5.4-mini` when you want a faster, lower-cost option for lighter coding

131 tasks or subagents. The `gpt-5.3-codex-spark` model is available in research

132 preview for ChatGPT Pro subscribers and is optimized for near-instant,

133 real-time coding iteration.

107 134 

108## Alternative models135## Alternative models

109 136 

134 161 

135The Codex CLI and IDE extension use the same `config.toml` [configuration file](https://developers.openai.com/codex/config-basic). To specify a model, add a `model` entry to your configuration file. If you don't specify a model, the Codex app, CLI, or IDE Extension defaults to a recommended model.162The Codex CLI and IDE extension use the same `config.toml` [configuration file](https://developers.openai.com/codex/config-basic). To specify a model, add a `model` entry to your configuration file. If you don't specify a model, the Codex app, CLI, or IDE Extension defaults to a recommended model.

136 163 

137```164```toml

138model = "gpt-5.4"165model = "gpt-5.5"

139```166```

140 167 

168If `gpt-5.5` isn't available in your account yet, use `gpt-5.4`.

169 

141### Choosing a different local model temporarily170### Choosing a different local model temporarily

142 171 

143In the Codex CLI, you can use the `/model` command during an active thread to change the model. In the IDE extension, you can use the model selector below the input box to choose your model.172In the Codex CLI, you can use the `/model` command during an active thread to change the model. In the IDE extension, you can use the model selector below the input box to choose your model.

145To start a new Codex CLI thread with a specific model or to specify the model for `codex exec` you can use the `--model`/`-m` flag:174To start a new Codex CLI thread with a specific model or to specify the model for `codex exec` you can use the `--model`/`-m` flag:

146 175 

147```bash176```bash

148codex -m gpt-5.4177codex -m gpt-5.5

149```178```

150 179 

151### Choosing your model for cloud tasks180### Choosing your model for cloud tasks

1# Remote connections1# Remote connections

2 2 

3SSH remote connections are currently in alpha. To enable them today, set3SSH remote connections are currently in alpha. To enable them today, set

4`remote_control = true` in the `[features]` table in `~/.codex/config.toml`.4 `remote_connections = true` in the `[features]` table in

5Availability, setup flows, and supported environments may change as the5 `~/.codex/config.toml`. Availability, setup flows, and supported environments

6feature improves.6 may change as the feature improves.

7 7 

8Remote connections let Codex work with projects that live on another8Remote connections let Codex work with projects that live on another

9SSH-accessible machine. Use them when the codebase, credentials, services, or9SSH-accessible machine. Use them when the codebase, credentials, services, or

48 48 

49```toml49```toml

50[features]50[features]

51remote_control = true51remote_connections = true

52```52```

53 53 

54Remote project threads run commands, read files, and write changes on the54Remote project threads run commands, read files, and write changes on the

5Codex offers the ability to increase the speed of the model for increased5Codex offers the ability to increase the speed of the model for increased

6credit consumption.6credit consumption.

7 7 

8Fast mode is currently supported on GPT-5.4. When enabled, speed is increased8Fast mode increases supported model speed by 1.5x and consumes credits at a

9by 1.5x and credits are consumed at a 2x rate.9higher rate than Standard mode. It currently supports GPT-5.5 and GPT-5.4,

10consuming credits at 2.5x the Standard rate for GPT-5.5 and 2x the Standard

11rate for GPT-5.4.

10 12 

11Use `/fast on`, `/fast off`, or `/fast status` in the CLI to change or inspect13Use `/fast on`, `/fast off`, or `/fast status` in the CLI to change or inspect

12the current setting. You can also persist the default with `service_tier = "fast"` plus `[features].fast_mode = true` in `config.toml`. Fast mode is14the current setting. You can also persist the default with `service_tier = "fast"` plus `[features].fast_mode = true` in `config.toml`. Fast mode is

20 22 

21## Codex-Spark23## Codex-Spark

22 24 

23GPT-5.3-Codex-Spark is a separate fast, less-capable Codex model optimized for near-instant, real-time coding iteration. Unlike fast mode, which speeds up GPT-5.4 at a higher credit rate,25GPT-5.3-Codex-Spark is a separate fast, less-capable Codex model optimized for

24Codex-Spark is its own model choice and has its own usage limits.26near-instant, real-time coding iteration. Unlike fast mode, which speeds up a

27supported model at a higher credit rate, Codex-Spark is its own model choice

28and has its own usage limits.

25 29 

26During research preview Codex-Spark is only available for ChatGPT Pro subscribers.30During research preview Codex-Spark is only available for ChatGPT Pro subscribers.