Command reference
neox launch flags, subcommands, and slash commands in the interactive UI.
Launch flags
| Flag | Short | Description |
|---|---|---|
--print | -p | Non-interactive: run this one prompt, print the result and exit |
--json | With -p, output JSON | |
--stream | With -p, stream output as it's generated | |
--output-schema <file> | Return structured output that follows a JSON Schema | |
--timeout <seconds> | With -p, stop after this long | |
--model <id> | -m | Pick a model; provider:model also works |
--provider <id> | Pick a provider | |
--dir <path> | -d | Working directory (also --workdir) |
--continue | -c | Continue the most recent session |
--resume [id] | -r | Continue a specific session; without an id, show a picker |
--no-session | Don't save the session | |
--yolo | Never ask before actions for this run (same as --dangerously-skip-permissions) | |
--lean | Lean mode: basic tools and a shorter system prompt, for this run only | |
--debug | Write more detailed logs | |
--debug-console | Print debug logs to the terminal | |
--help | -h | Help |
--version | -v | Version |
Subcommands
| Command | Description |
|---|---|
neox login | Sign in; add --device-code where there's no browser |
neox logout | Sign out |
neox whoami | Current account |
neox usage | Usage and quota |
neox device | Signed-in devices |
neox upgrade | Open the subscription page |
neox model ls | List available models (--source=cloud, byok or all) |
neox provider ls / add / test [id] | Bring-your-own-key providers: list, add, test the connection |
neox mcp ... | Manage MCP servers — see MCP |
neox skill ... | Skills: list, install <url or path>, preview, update, uninstall, trust / untrust |
neox browser ls / replay <name> | List / replay recorded browser scripts; replay --all replays them all |
neox daemon ... | Background service — see Daemon |
neox update | Check for updates and update |
neox uninstall | Uninstall |
neox "do X" (when it isn't a subcommand name) is the same as neox -p "do X".
Slash commands in the interactive UI
Type / to open the command menu, then type to filter.
Basics
| Command | Description |
|---|---|
/help | Help |
/setup | Setup wizard: service, model, web search |
/clear | Clear the conversation context (text already on screen stays) |
/exit | Quit (or /quit) |
/feedback | Send feedback |
Account
| Command | Description |
|---|---|
/login / /logout | Sign in / out |
/whoami | Current account |
/usage | Usage and quota |
/device | Signed-in devices |
/upgrade | Upgrade your plan (opens the web page) |
Modes
| Command | Description |
|---|---|
/mode | agent does the work / ask just talks |
/thinking | Deep thinking on / off |
/approval | Whether to ask before actions: auto / manual / yolo |
/sandbox | Sandbox: limit which files commands can change |
/theme | Color theme |
Models
| Command | Description |
|---|---|
/model | Switch model |
/provider | Providers and API keys |
/effort | Reasoning effort |
/output-style | Reply style: standard / concise / detailed / code only / custom |
/diagnostics | Request diagnostics such as cache hits |
Sessions
| Command | Description |
|---|---|
/resume | Continue a previous session |
/session | ls / new / info / export / clear |
/rollback | Roll file changes back to a checkpoint — see Checkpoints |
/checkpoint | Checkpoint menu: create manually, list, clean up |
/compact | Compress the conversation to free up context |
Tools
| Command | Description |
|---|---|
/attach | Attach a file, image or URL to the next message |
/workspace | Switch the working directory |
/websearch | Web search |
/mcp | MCP servers |
/skills | Skills |
/context | How much context is in use |
/memory | Memory |
/init | Read the project and generate project memory (.neox/project.md) |
/remote | Phone / remote access |
/ps | Background commands (/kill to stop one) |
Stats & config
| Command | Description |
|---|---|
/stats | Current session: directory, model, account, mode |
/statistic | Token usage statistics |
/language | Interface language |
/notify | Completion alerts |
/tts | Read replies aloud: Edge / OpenAI / off |
/index | Code index |
/cleanup | Disk usage |
/experimental | Experimental features |
/update | Check for updates |
Installed skills can also be called directly as /<skill name> — see Skills.

