Slash commands

Brain works fine in plain English. Slash commands are for cases where you want explicit control: change the underlying LLM, see what’s available, get help.

Help & discovery

Command	What it does
`/help`	List every command + a one-line description.
`/provider list`	Show every provider (Gemini, OpenAI, Mistral, Claude) + the model id at each tier (small / medium / large).
`/model list`	Same matrix, grouped by tier × provider.

Switch provider (persistent)

/provider sets the LLM that handles all your turns going forward, until you change it again.

Command	Effect
`/provider`	Show current provider.
`/provider gemini`	Use Google Gemini.
`/provider openai`	Use OpenAI (GPT).
`/provider mistral`	Use Mistral.
`/provider claude`	Use Anthropic Claude.
`/provider reset`	Back to default (Gemini).

Natural language works too: “use mistral as provider”, “switch to claude”.

Switch tier (persistent)

/model sets the size/cost tier within whichever provider is active.

Command	Effect
`/model small`	Cheapest, fastest.
`/model medium`	Balanced (default).
`/model large`	Most capable.
`/model reset`	Back to default.

Natural language: “use the small model”, “utilise le modèle large”.

Per-turn override (one-shot)

Prefix a message with a provider or tier name to use it for this turn only — your persistent settings stay untouched.

/openai draft a follow-up to Acme
/claude write a long-form analysis of last quarter
/sonnet schedule a meeting with Pierre
/haiku what time is it in Tokyo

Recognised per-turn prefixes: /gemini, /openai, /mistral, /claude, /anthropic, /code, /opus, /sonnet, /haiku, /fast, /think-hard.

Skills & discovery

Command	What it does
`/help <name>`	A compact card for one skill or scheduled task — instant, no LLM turn.
`/skills`	List your saved skills.
`/create-skill`	Build a skill by chatting — Brain interviews you, drafts it, and self-reviews.
`/review-skill <name>`	Brain critiques and improves an existing skill; you approve.
`/share-skill <name>`	Publish a skill to your team (everyone on your verified email domain).
`/org-skills`	Browse skills your teammates have shared.
`/org-skill <name>`	Preview a shared skill and import it into your vault.
`/unshare-skill <name>`	Withdraw a skill you published to your team.
`/skill <name>`	Open the guided editor for one of your skills (bare `/skill` → use `/skills`).
`/skill-versions`	`list` / `show` / `diff` / `restore` a skill’s earlier versions.

See Skills for the full how-to. Team skills are also browsable at org.nitroxbrain.com.

Scheduled tasks

Command	What it does
`/tasks`	List every scheduled task you have (morning brief, digests, …).
`/task`	Start a guided interview to create a new scheduled task.
`/task <name>`	Edit that scheduled task — or create it if it doesn’t exist yet.

Session

Command	What it does
`/clear`	Start a fresh conversation. Your long-term memory is untouched — only the current thread resets.
`/dual`	Show dual-mode status; `/dual on` / `/dual off` toggles it. Each turn then runs through two LLM providers, reconciled into one answer — a built-in second opinion. Paid plan + enablement on your account required; roughly doubles cost and latency per turn.

French aliases

/aide mirrors /help and /taches mirrors /tasks, rendering the output in French.

What’s not a slash command

Asking Brain to do something — just write it in plain English.
Switching languages — Brain detects the language of your message; no command needed.
Reading a file — just paste / drop / mention it.