@chinmaymk/aikit / OpenAIOptions
Interface: OpenAIOptions
Defined in: types.ts:451
OpenAI Chat Completions API configuration and generation options (default OpenAI implementation). These can be provided at construction time or generation time. Generation time options will override construction time options.
Extends
Properties
model?
optionalmodel:string
Defined in: types.ts:305
The specific model you want to use. e.g., 'gpt-4o' or 'claude-3-5-sonnet-20240620'.
Inherited from
maxOutputTokens?
optionalmaxOutputTokens:number
Defined in: types.ts:307
The maximum number of output tokens to generate. Don't want it to ramble on forever, do you?
Inherited from
ProviderOptions.maxOutputTokens
temperature?
optionaltemperature:number
Defined in: types.ts:313
The sampling temperature. Higher values (e.g., 0.8) make the output more random, while lower values (e.g., 0.2) make it more focused and deterministic. A bit like adjusting the chaos knob.
Inherited from
topP?
optionaltopP:number
Defined in: types.ts:318
Top-p sampling. It's a way to control the randomness of the output by only considering the most likely tokens. It's like telling the AI to only pick from the top of the deck.
Inherited from
topK?
optionaltopK:number
Defined in: types.ts:323
Top-k sampling. Similar to top-p, but it considers a fixed number of top tokens. Not all providers support this, because life isn't fair.
Inherited from
stopSequences?
optionalstopSequences:string[]
Defined in: types.ts:325
A list of sequences that will stop the generation. A safe word, if you will.
Inherited from
tools?
optionaltools:Tool[]
Defined in: types.ts:327
The list of tools you're making available to the model.
Inherited from
toolChoice?
optionaltoolChoice: {name:string; } |"auto"|"required"|"none"
Defined in: types.ts:335
How the model should choose which tool to use. 'auto': The model decides. 'required': The model must use a tool. 'none': The model can't use any tools. { name: 'my_tool' }: Force the model to use a specific tool.
Inherited from
apiKey?
optionalapiKey:string
Defined in: types.ts:344
API key for authentication with the provider.
Inherited from
baseURL?
optionalbaseURL:string
Defined in: types.ts:346
Custom base URL for the API endpoint.
Inherited from
timeout?
optionaltimeout:number
Defined in: types.ts:348
Request timeout in milliseconds.
Inherited from
maxRetries?
optionalmaxRetries:number
Defined in: types.ts:350
Maximum number of retry attempts for failed requests.
Inherited from
mutateHeaders()?
optionalmutateHeaders: (headers) =>void
Defined in: types.ts:356
A function that allows you to modify the headers before a request is sent. This is useful for adding custom headers or modifying existing ones.
Parameters
headers
Record<string, string>
The original headers object to mutate directly.
Returns
void
Inherited from
organization?
optionalorganization:string
Defined in: types.ts:453
Your OpenAI organization ID. For when you're part of a fancy club.
project?
optionalproject:string
Defined in: types.ts:455
Your OpenAI project ID. For even fancier clubs.
presencePenalty?
optionalpresencePenalty:number
Defined in: types.ts:461
Presence penalty. Positive values penalize new tokens based on whether they appear in the text so far, increasing the model's likelihood to talk about new topics. Basically, it discourages repetition.
frequencyPenalty?
optionalfrequencyPenalty:number
Defined in: types.ts:467
Frequency penalty. Positive values penalize new tokens based on their existing frequency in the text so far, decreasing the model's likelihood to repeat the same line verbatim. Stop me if you've heard this one before.
user?
optionaluser:string
Defined in: types.ts:471
A stable identifier for your end-users.
logprobs?
optionallogprobs:boolean
Defined in: types.ts:477
Whether to return log probabilities of the output tokens or not. If true, returns the log probabilities of each output token returned in the content of message. This feature is available on gpt-4.1, gpt-4o, gpt-4o-mini, gpt-3.5-turbo, and other supported models.
topLogprobs?
optionaltopLogprobs:number
Defined in: types.ts:482
An integer between 0 and 20 specifying the number of most likely tokens to return at each token position, each with an associated log probability. logprobs must be set to true if this parameter is used.
seed?
optionalseed:number
Defined in: types.ts:489
This feature is in Beta. If specified, our system will make a best effort to sample deterministically, such that repeated requests with the same seed and parameters should return the same result. Determinism is not guaranteed, and you should refer to the system_fingerprint response parameter to monitor changes in the backend.
responseFormat?
optionalresponseFormat:object
Defined in: types.ts:499
An object specifying the format that the model must output. Compatible with GPT-4.1, GPT-4o, GPT-4o-mini, GPT-3.5 Turbo, and all GPT-4 Turbo models newer than gpt-4-turbo-2024-04-09. Setting to { "type": "json_object" } enables JSON mode, which guarantees the message the model generates is valid JSON. Important: when using JSON mode, you must also instruct the model to produce JSON yourself via a system or user message. Without this, the model may generate an unending stream of whitespace until the generation reaches the token limit, resulting in a long-running and seemingly "stuck" request. Also note that the message content may be partially cut off if finish_reason="length", which indicates the generation exceeded max_tokens or the conversation exceeded the max context length.
type
type:
"text"|"json_object"|"json_schema"
json_schema?
optionaljson_schema:object
json_schema.name
name:
string
json_schema.description?
optionaldescription:string
json_schema.schema?
optionalschema:Record<string,unknown>
json_schema.strict?
optionalstrict:boolean
logitBias?
optionallogitBias:Record<string,number>
Defined in: types.ts:515
Modify the likelihood of specified tokens appearing in the completion. Accepts a JSON object that maps tokens (specified by their token ID in the tokenizer) to an associated bias value from -100 to 100. Mathematically, the bias is added to the logits generated by the model prior to sampling. The exact effect will vary per model, but values between -1 and 1 should decrease or increase likelihood of selection; values like -100 or 100 should result in a ban or exclusive selection of the relevant token.
n?
optionaln:number
Defined in: types.ts:521
How many chat completion choices to generate for each input message. Note that you will be charged based on the number of generated tokens across all of the choices. Keep n as 1 to minimize costs.
includeUsage?
optionalincludeUsage:boolean
Defined in: types.ts:526
Whether to include usage information (token counts and timing) in the response. When true, usage data will be included in the final stream chunk.
parallelToolCalls?
optionalparallelToolCalls:boolean
Defined in: types.ts:530
Whether to allow the model to run tool calls in parallel.
reasoning?
optionalreasoning:object
Defined in: types.ts:536
Configuration options for reasoning models (o-series models only). Controls the reasoning effort level for enhanced problem-solving capabilities. When enabled, the reasoning process is included in the response.
effort?
optionaleffort:"low"|"medium"|"high"
modalities?
optionalmodalities: ("text"|"audio")[]
Defined in: types.ts:543
Output types that you would like the model to generate. Most models are capable of generating text, which is the default. The gpt-4o-audio-preview model can also be used to generate audio.
audio?
optionalaudio:object
Defined in: types.ts:547
Parameters for audio output. Required when audio output is requested with modalities: ["audio"].
format
format:
"wav"|"mp3"|"flac"|"opus"|"pcm16"
Specifies the output audio format. Must be one of wav, mp3, flac, opus, or pcm16.
voice
voice:
"alloy"|"ash"|"ballad"|"coral"|"echo"|"fable"|"nova"|"onyx"|"sage"|"shimmer"
The voice the model uses to respond.
maxCompletionTokens?
optionalmaxCompletionTokens:number
Defined in: types.ts:567
The maximum number of tokens that can be generated in the chat completion, including visible output tokens and reasoning tokens. This replaces the deprecated max_tokens parameter.
prediction?
optionalprediction:object
Defined in: types.ts:571
Configuration for a Predicted Output, which can greatly improve response times when large parts of the model response are known ahead of time.
type
type:
"content"
The type of predicted content. Currently only 'content' is supported.
content
content:
string|object[]
The content that should be matched when generating a model response.
webSearchOptions?
optionalwebSearchOptions:object
Defined in: types.ts:580
Web search tool configuration for searching the web for relevant results to use in a response.
searchContextSize?
optionalsearchContextSize:"low"|"medium"|"high"
High level guidance for the amount of context window space to use for the search.
userLocation?
optionaluserLocation:object
Approximate location parameters for the search.
userLocation.type
type:
"approximate"
The type of location approximation. Always 'approximate'.
userLocation.approximate
approximate:
Record<string,unknown>
Approximate location parameters for the search.