De interface van VDS Renaissance is in meerdere talen beschikbaar. De officiële documentatie wordt in het Engels onderhouden om één consistente en altijd actuele referentie te garanderen.

@AI

Windows ✓Linux ⚠macOS ⚠

[VDS7]

Syntax

@AI(<prompt>)

@AI(STATUS)

@AI(TOKENS)

@AI(ERROR)

@AI(MODEL)

@AI(CHUNK)

@AI(STREAMING)

@AI(RESPONSE)

Description

This function is the call that talks to the AI model. Used with a prompt, it sends that prompt (together with the current conversation history) and returns the model's reply as text. Configure the provider and model first with the AI command.

%reply = @ai(Summarise this in one sentence: %text)

A prompt containing commas must be passed through a variable, because a comma is a parameter separator.

After a call, the following selectors return information about it:

@AI(STATUS) The HTTP status of the last call (200 on success, or an error code).
@AI(TOKENS) The tokens consumed, input and output, joined by the current field separator.
@AI(ERROR) The error message for the last call, or empty on success.
@AI(MODEL) The model actually used for the last call.
@AI(CHUNK) The next fragment of a streamed reply (empty if nothing new). See AI streaming.
@AI(STREAMING) 1 while a streamed reply is still arriving, 0 once it is complete.
@AI(RESPONSE) The text of a streamed reply assembled so far (complete once streaming ends).

The last three are used with ai stream for progressive replies.

OK

When called with a prompt, OK is set false if the call failed; @AI(error) then gives the reason (for example a missing API key, an unknown model, or an HTTP error).

Example

ai config,anthropic,claude-opus-4-8
%answer = @ai(Name the capital of France.)
if @ok()
  info %answer @cr()(@ai(tokens) tokens)
else
  info Error: @ai(error)
end

See also