@markprompt/core

@markprompt/core

@markprompt/core is the core library for Markprompt, a conversational AI component for your website, trained on your data.

It contains core functionality for Markprompt and allows you to build abstractions on top of it.

Installation

npm install @markprompt/core

In browsers with esm.sh:

<script type="module">
  import {
    submitChat,
    submitSearchQuery,
    submitFeedback,
  } from 'https://esm.sh/@markprompt/core';
</script>

Usage

import { submitChat } from '@markprompt/core';

for await (const chunk of submitChat(
  [{ content: 'What is Markprompt?', role: 'user' }],
  'YOUR-PROJECT-KEY',
  {
    model: 'gpt-4-turbo-preview',
    systemPrompt: 'You are a helpful AI assistant'
  }
)) {
  console.debug(chunk);
}

API

submitChat(messages: ChatMessage[], projectKey: string, options?)

Submit a prompt to the Markprompt Completions API.

Arguments

• messages (ChatMessage[]): Chat messages to submit to the model
• projectKey (string): Project key for the project
• options (SubmitChatOptions): Optional parameters

Options

All options are optional.

• apiUrl (string): URL at which to fetch completions
• conversationId (string): Conversation ID
• iDontKnowMessage (string): Message returned when the model does not have an answer
• model (OpenAIModelId): The OpenAI model to use
• systemPrompt (string): The prompt template
• temperature (number): The model temperature
• topP (number): The model top P
• frequencyPenalty (number): The model frequency penalty
• presencePenalty (number): The model present penalty
• maxTokens (number): The max number of tokens to include in the response
• sectionsMatchCount (number): The number of sections to include in the prompt context
• sectionsMatchThreshold (number): The similarity threshold between the
• signal (AbortSignal): AbortController signal
• tools: (OpenAI.ChatCompletionTool[]): A list of tools the model may call
• tool_choice: (OpenAI.ChatCompletionToolChoiceOption): Controls which (if any) function is called by the model

Returns

A promise that resolves when the response is fully handled.

submitSearchQuery(query, projectKey, options?)

Submit a search query to the Markprompt Search API.

Arguments

• query (string): Search query
• projectKey (string): Project key for the project
• options (object): Optional parameters

Options

• apiUrl (string): URL at which to fetch search results
• limit (number): Maximum amount of results to return
• signal (AbortSignal): AbortController signal

Returns

A list of search results.

submitFeedback(feedback, projectKey, options?)

Submit feedback to the Markprompt Feedback API about a specific prompt.

Arguments

• feedback (object): Feedback to submit
• feedback.feedback (object): Feedback data
• feedback.feedback.vote ("1" | "-1" | "escalated"): Vote
• feedback.promptId (string): Prompt ID
• projectKey (string): Project key for the project
• options (object): Optional parameters
• options.apiUrl (string): URL at which to post feedback
• options.onFeedbackSubmitted (function): Callback function when feedback is submitted
• options.signal (AbortSignal): AbortController signal

Returns

A promise that resolves when the feedback is submitted. Has no return value.

Community

• X
• Discord

Authors

This library is created by the team behind Markprompt (@markprompt).

License

MIT © Markprompt