Assistant commands

AlayaCode's web chat (/chat) and most integrated CLIs recognise a set of slash commands — quick ways to check usage, switch models, reset context — so you don't need to hop into the console.

/help

Lists every available command with a one-liner. Use this whenever you're not sure what's next.

/usage

Current subscription-cycle usage:

• Spent / total quota (in tokens)
• Days remaining
• Current rate-limit tier (RPM / TPM)
• Headroom before throttling kicks in

/models

Lists the model IDs available under your subscription. Equivalent to GET /v1/models, formatted for humans.

/switch <model_id>

Switch the model used for this session, e.g.

/switch glm-5.1

Affects only the active session — not your account-level default.

/reset

Clears the in-session message history. Equivalent to "start a new conversation". It does not delete the server-side session log — audit and billing records remain.

/system <prompt>

Set a system prompt for the active session. Subsequent requests carry it. Calling /system again replaces it; /system followed by an empty string clears it.

/temp <number>

Adjust sampling temperature for this session, e.g. /temp 0.3. Range 0-2 (OpenAI) or 0-1 (Anthropic).

/exit

Exits the session (CLI) or closes the chat panel (web). Doesn't change subscription state.

Slash commands inside the SDK?

These are chat-UI sugar — they're not forwarded to the model, so SDK calls to /v1/chat/completions ignore them. From SDKs, use plain parameters: switch models via the model field, reset by reconstructing the messages array.

Command set varies by subscription tier (higher tiers expose extras like /audit). /help always shows the live set for your key — trust that as the source of truth.