Generative AI

Functions

Conversation Add Message

Adds a message to a Conversation, that the language model will begin replying to. You can receive the reply one piece at a time by calling conversation_get_reply_piece(conversation c) in a loop

Parameters:

Name	Type	Description
c	`Conversation`	The `Conversation` object to check
message	`String`	The user message to add to the conversation - the language model will reply to this

Signatures:

void conversation_add_message(conversation c, const string &message)

public void Conversation.AddMessage(string message);
public static void SplashKit.ConversationAddMessage(Conversation c, string message);

def conversation_add_message(c, message):

procedure ConversationAddMessage(c: Conversation; const message: String)

Conversation Get Reply

Returns a reply from a Conversation, without any related thoughts.

Parameters:

Name	Type	Description
conv	`Conversation`	The `Conversation` to recieve the reply from

Return Type: String

Returns: The response from the model

Signatures:

string conversation_get_reply(conversation conv)

public string Conversation.GetReply();
public static string SplashKit.ConversationGetReply(Conversation conv);

def conversation_get_reply(conv):

function ConversationGetReply(conv: Conversation): String

Conversation Get Reply

Returns a reply from a Conversation, with the ability to indicate if thoughts should be included.

Parameters:

Name	Type	Description
conv	`Conversation`	The `Conversation` to recieve the reply from
with_thoughts	`Boolean`	A boolean to indicate if thoughts should be included in the reply

Return Type: String

Returns: The response from the model

Signatures:

string conversation_get_reply(conversation conv, bool with_thoughts)

public string Conversation.GetReply(bool withThoughts);
public static string SplashKit.ConversationGetReply(Conversation conv, bool withThoughts);

def conversation_get_reply_with_thoughts(conv, with_thoughts):

function ConversationGetReply(conv: Conversation; withThoughts: Boolean): String

Conversation Get Reply Piece

Returns a single piece of a reply (generally one word at a time) from the Conversation You can use a loop while checking Conversation Is Replying to retrieve the reply as it generates

Parameters:

Name	Type	Description
c	`Conversation`	The `Conversation` object to recieve the reply from

Return Type: String

Returns: Returns a small piece of the reply (generally 1 word or less)

Signatures:

string conversation_get_reply_piece(conversation c)

public string Conversation.GetReplyPiece();
public static string SplashKit.ConversationGetReplyPiece(Conversation c);

def conversation_get_reply_piece(c):

function ConversationGetReplyPiece(c: Conversation): String

Conversation Is Replying

Checks if a language model is currently generating a reply within a Conversation. If so, you can continue to receive the message with conversation_get_reply_piece(conversation c)

Parameters:

Name	Type	Description
c	`Conversation`	The `Conversation` object to check

Return Type: Boolean

Returns: Returns whether the language model is still generating a reply

Signatures:

bool conversation_is_replying(conversation c)

public bool Conversation.IsReplying();
public static bool SplashKit.ConversationIsReplying(Conversation c);

def conversation_is_replying(c):

function ConversationIsReplying(c: Conversation): Boolean

Conversation Is Thinking

Checks if a language model is currently “thinking” while generating a reply within a Conversation. You can use this to filter out the “thoughts” and display them differently (or hide them entirely)

Parameters:

Name	Type	Description
c	`Conversation`	The `Conversation` object to check

Return Type: Boolean

Returns: Returns whether the language model is currently thinking while generating a reply

Signatures:

bool conversation_is_thinking(conversation c)

public bool Conversation.IsThinking();
public static bool SplashKit.ConversationIsThinking(Conversation c);

def conversation_is_thinking(c):

function ConversationIsThinking(c: Conversation): Boolean

Creates a new Conversation object, that uses the default language model. The Conversation object can have messages added to it, and responses streamed back from it via the other Conversation functions and procedures

Return Type: Conversation

Returns: Returns a new Conversation object.

Signatures:

conversation create_conversation()

public static Conversation SplashKit.CreateConversation();
public Conversation();

def create_conversation():

function CreateConversation(): Conversation

Create Conversation

Creates a new Conversation object, that uses a chosen language model. The Conversation object can have messages added to it, and responses streamed back from it via the other Conversation functions and procedures

Parameters:

Name	Type	Description
model	`Language Model`	The language model to use

Return Type: Conversation

Returns: Returns a new Conversation object.

Signatures:

conversation create_conversation(language_model model)

public static Conversation SplashKit.CreateConversation(LanguageModel model);
public Conversation(LanguageModel model);

def create_conversation_with_model(model):

function CreateConversation(model: LanguageModel): Conversation

Free All Conversations

Releases all of the Conversation objects which have been loaded.

Signatures:

void free_all_conversations()

public static void GenerativeAi.FreeAll();
public static void SplashKit.FreeAllConversations();

def free_all_conversations():

procedure FreeAllConversations()

Free Conversation

Frees the resources associated with the Conversation object.

Parameters:

Name	Type	Description
c	`Conversation`	The `Conversation` object whose resources should be released.

Signatures:

void free_conversation(conversation c)

public void Conversation.Free();
public static void SplashKit.FreeConversation(Conversation c);

def free_conversation(c):

procedure FreeConversation(c: Conversation)

Generate Reply

Generates a reply to a textual prompt by a language model The language model will respond to the textual prompt in a chat style format. It will follow instructions and answer questions. Instruct or Thinking models are recommended. Base models likely won’t output sensible results.

Parameters:

Name	Type	Description
model	`Language Model`	The language model to use
prompt	`String`	The prompt for the language model to reply to.

Return Type: String

Returns: The generated reply.

Signatures:

string generate_reply(language_model model, string prompt)

public static string SplashKit.GenerateReply(LanguageModel model, string prompt);

def generate_reply_with_model(model, prompt):

function GenerateReply(model: LanguageModel; prompt: String): String

Generate Reply

Parameters:

Name	Type	Description
prompt	`String`	The prompt for the language model to reply to.

Return Type: String

Returns: The generated reply.

Signatures:

string generate_reply(string prompt)

public static string SplashKit.GenerateReply(string prompt);

def generate_reply(prompt):

function GenerateReply(prompt: String): String

Generate Text

Generates text that continues from a prompt, with a maximum of 125 tokens. The language model will continue predicting text based on patterns in the prompt - it will not directly follow instructions or answer questions. Base models are recommended; Instruct and Thinking models may work.

Parameters:

Name	Type	Description
model	`Language Model`	The language model to use
text	`String`	The input text for the language model to continue.

Return Type: String

Returns: The generated reply.

Signatures:

string generate_text(language_model model, string text)

public static string SplashKit.GenerateText(LanguageModel model, string text);

def generate_text_with_model(model, text):

function GenerateText(model: LanguageModel; text: String): String

Generate Text

Parameters:

Name	Type	Description
model	`Language Model`	The language model to use
text	`String`	The input text for the language model to continue.
max_tokens	`Integer`	The maximum tokens used in response - determining the length of the output and the time taken. Keep this small for reasonable execution times.

Return Type: String

Returns: The generated reply.

Signatures:

string generate_text(language_model model, string text, int max_tokens)

public static string SplashKit.GenerateText(LanguageModel model, string text, int maxTokens);

def generate_text_with_model_and_tokens(model, text, max_tokens):

function GenerateText(model: LanguageModel; text: String; maxTokens: Integer): String

Generate Text

Generates text that continues from a prompt - with default of 125 tokens. The language model will continue predicting text based on patterns in the prompt - it will not directly follow instructions or answer questions. Base models are recommended; Instruct and Thinking models may work.

Parameters:

Name	Type	Description
text	`String`	The input text for the language model to continue.

Return Type: String

Returns: The generated reply.

Signatures:

string generate_text(string text)

public static string SplashKit.GenerateText(string text);

def generate_text(text):

function GenerateText(text: String): String

Generate Text

Generates text that continues from a prompt. The language model will continue predicting text based on patterns in the prompt - it will not directly follow instructions or answer questions. Base models are recommended; Instruct and Thinking models may work.

Parameters:

Name	Type	Description
text	`String`	The input text for the language model to continue.
max_tokens	`Integer`	The maximum tokens used in response - determining the length of the output and the time taken. Keep this small for reasonable execution times.

Return Type: String

Returns: The generated reply.

Signatures:

string generate_text(string text, int max_tokens)

public static string SplashKit.GenerateText(string text, int maxTokens);

def generate_text_with_tokens(text, max_tokens):

function GenerateText(text: String; maxTokens: Integer): String

Types

Conversation

The Conversation type is used to refer to conversations between the user and a language model. You can use it to send messages to the language model, and stream responses back.

All Conversation objects are:

created with create_conversation(), create_conversation(language_model model)
and must be released using free_conversation() (to release a specific Conversation object) or free_all_conversation() (to release all created Conversation objects).

Language Model

Language Models: Choose between different language models to trade off speed and intelligence Each model is scaled to fit within 1~2GB and will be automatically downloaded when needed - feel free to try them out!

Constant	Description
QWEN3_0_6B_BASE	Qwen3 0.6B Base model - small, extremely fast and good for text completion. Very limited world knowledge.
QWEN3_0_6B_INSTRUCT	Qwen3 0.6B Instruct model (default) - small, extremely fast and can follow simple instructions. Very limited world knowledge.
QWEN3_0_6B_THINKING	Qwen3 0.6B Thinking model - small, extremely fast and can follow more specific instructions, but has a short delay before starting to reply. Very limited world knowledge.
QWEN3_1_7B_BASE	Qwen3 1.7B Base model - decently fast and good for text completion. Limited world knowledge.
QWEN3_1_7B_INSTRUCT	Qwen3 1.7B Instruct model - decently fast and can follow instructions. Limited world knowledge.
QWEN3_1_7B_THINKING	Qwen3 1.7B Thinking model - decently fast and can follow more difficult instructions, but has a delay before starting to reply. Limited world knowledge.
QWEN3_4B_BASE	Qwen3 4B Base model - slower but excellent for text completion/pattern based completion
QWEN3_4B_INSTRUCT	Qwen3 4B Instruct model - slower but can follow complex instructions
QWEN3_4B_THINKING	Qwen3 4B Thinking model - slower but can follow complex and specific instructions, but has a potentially long delay before starting to reply
GEMMA3_270M_BASE	Gemma3 270M Base model - tiny, extremely fast, and good for text completion. Very limited world knowledge.
GEMMA3_270M_INSTRUCT	Gemma3 270M Instruct model - tiny, extremely fast, and good for very simple instructions. Very limited world knowledge.
GEMMA3_1B_BASE	Gemma3 1B Base model - fast and good for text completion. Has decent world knowledge and multilingual abilities.
GEMMA3_1B_INSTRUCT	Gemma3 1B Instruct model - fast and can follow instructions. Has decent world knowledge and multilingual abilities.
GEMMA3_4B_BASE	Gemma3 4B Base model - slower but good for text completion/pattern based completion. Has decent world knowledge and multilingual abilities.
GEMMA3_4B_INSTRUCT	Gemma3 4B Instruct model - slower but can follow complex instructions. Has decent world knowledge and multilingual abilities.

Constant	Description
LanguageModel.Qwen306BBase	Qwen3 0.6B Base model - small, extremely fast and good for text completion. Very limited world knowledge.
LanguageModel.Qwen306BInstruct	Qwen3 0.6B Instruct model (default) - small, extremely fast and can follow simple instructions. Very limited world knowledge.
LanguageModel.Qwen306BThinking	Qwen3 0.6B Thinking model - small, extremely fast and can follow more specific instructions, but has a short delay before starting to reply. Very limited world knowledge.
LanguageModel.Qwen317BBase	Qwen3 1.7B Base model - decently fast and good for text completion. Limited world knowledge.
LanguageModel.Qwen317BInstruct	Qwen3 1.7B Instruct model - decently fast and can follow instructions. Limited world knowledge.
LanguageModel.Qwen317BThinking	Qwen3 1.7B Thinking model - decently fast and can follow more difficult instructions, but has a delay before starting to reply. Limited world knowledge.
LanguageModel.Qwen34BBase	Qwen3 4B Base model - slower but excellent for text completion/pattern based completion
LanguageModel.Qwen34BInstruct	Qwen3 4B Instruct model - slower but can follow complex instructions
LanguageModel.Qwen34BThinking	Qwen3 4B Thinking model - slower but can follow complex and specific instructions, but has a potentially long delay before starting to reply
LanguageModel.Gemma3270mBase	Gemma3 270M Base model - tiny, extremely fast, and good for text completion. Very limited world knowledge.
LanguageModel.Gemma3270mInstruct	Gemma3 270M Instruct model - tiny, extremely fast, and good for very simple instructions. Very limited world knowledge.
LanguageModel.Gemma31BBase	Gemma3 1B Base model - fast and good for text completion. Has decent world knowledge and multilingual abilities.
LanguageModel.Gemma31BInstruct	Gemma3 1B Instruct model - fast and can follow instructions. Has decent world knowledge and multilingual abilities.
LanguageModel.Gemma34BBase	Gemma3 4B Base model - slower but good for text completion/pattern based completion. Has decent world knowledge and multilingual abilities.
LanguageModel.Gemma34BInstruct	Gemma3 4B Instruct model - slower but can follow complex instructions. Has decent world knowledge and multilingual abilities.

Constant	Description
LanguageModel.qwen3_0_6b_base	Qwen3 0.6B Base model - small, extremely fast and good for text completion. Very limited world knowledge.
LanguageModel.qwen3_0_6b_instruct	Qwen3 0.6B Instruct model (default) - small, extremely fast and can follow simple instructions. Very limited world knowledge.
LanguageModel.qwen3_0_6b_thinking	Qwen3 0.6B Thinking model - small, extremely fast and can follow more specific instructions, but has a short delay before starting to reply. Very limited world knowledge.
LanguageModel.qwen3_1_7b_base	Qwen3 1.7B Base model - decently fast and good for text completion. Limited world knowledge.
LanguageModel.qwen3_1_7b_instruct	Qwen3 1.7B Instruct model - decently fast and can follow instructions. Limited world knowledge.
LanguageModel.qwen3_1_7b_thinking	Qwen3 1.7B Thinking model - decently fast and can follow more difficult instructions, but has a delay before starting to reply. Limited world knowledge.
LanguageModel.qwen3_4b_base	Qwen3 4B Base model - slower but excellent for text completion/pattern based completion
LanguageModel.qwen3_4b_instruct	Qwen3 4B Instruct model - slower but can follow complex instructions
LanguageModel.qwen3_4b_thinking	Qwen3 4B Thinking model - slower but can follow complex and specific instructions, but has a potentially long delay before starting to reply
LanguageModel.gemma3_270m_base	Gemma3 270M Base model - tiny, extremely fast, and good for text completion. Very limited world knowledge.
LanguageModel.gemma3_270m_instruct	Gemma3 270M Instruct model - tiny, extremely fast, and good for very simple instructions. Very limited world knowledge.
LanguageModel.gemma3_1b_base	Gemma3 1B Base model - fast and good for text completion. Has decent world knowledge and multilingual abilities.
LanguageModel.gemma3_1b_instruct	Gemma3 1B Instruct model - fast and can follow instructions. Has decent world knowledge and multilingual abilities.
LanguageModel.gemma3_4b_base	Gemma3 4B Base model - slower but good for text completion/pattern based completion. Has decent world knowledge and multilingual abilities.
LanguageModel.gemma3_4b_instruct	Gemma3 4B Instruct model - slower but can follow complex instructions. Has decent world knowledge and multilingual abilities.

Constant	Description
LanguageModel.Qwen306BBase	Qwen3 0.6B Base model - small, extremely fast and good for text completion. Very limited world knowledge.
LanguageModel.Qwen306BInstruct	Qwen3 0.6B Instruct model (default) - small, extremely fast and can follow simple instructions. Very limited world knowledge.
LanguageModel.Qwen306BThinking	Qwen3 0.6B Thinking model - small, extremely fast and can follow more specific instructions, but has a short delay before starting to reply. Very limited world knowledge.
LanguageModel.Qwen317BBase	Qwen3 1.7B Base model - decently fast and good for text completion. Limited world knowledge.
LanguageModel.Qwen317BInstruct	Qwen3 1.7B Instruct model - decently fast and can follow instructions. Limited world knowledge.
LanguageModel.Qwen317BThinking	Qwen3 1.7B Thinking model - decently fast and can follow more difficult instructions, but has a delay before starting to reply. Limited world knowledge.
LanguageModel.Qwen34BBase	Qwen3 4B Base model - slower but excellent for text completion/pattern based completion
LanguageModel.Qwen34BInstruct	Qwen3 4B Instruct model - slower but can follow complex instructions
LanguageModel.Qwen34BThinking	Qwen3 4B Thinking model - slower but can follow complex and specific instructions, but has a potentially long delay before starting to reply
LanguageModel.Gemma3270mBase	Gemma3 270M Base model - tiny, extremely fast, and good for text completion. Very limited world knowledge.
LanguageModel.Gemma3270mInstruct	Gemma3 270M Instruct model - tiny, extremely fast, and good for very simple instructions. Very limited world knowledge.
LanguageModel.Gemma31BBase	Gemma3 1B Base model - fast and good for text completion. Has decent world knowledge and multilingual abilities.
LanguageModel.Gemma31BInstruct	Gemma3 1B Instruct model - fast and can follow instructions. Has decent world knowledge and multilingual abilities.
LanguageModel.Gemma34BBase	Gemma3 4B Base model - slower but good for text completion/pattern based completion. Has decent world knowledge and multilingual abilities.
LanguageModel.Gemma34BInstruct	Gemma3 4B Instruct model - slower but can follow complex instructions. Has decent world knowledge and multilingual abilities.