ai-fetcher/dist/esm/libs/claude.js

A Node.js package that provides integration with popular language models. It is designed to facilitate easy and efficient ai-fetching tasks for your application.

import axios from "axios";
// The Claude Agent class
export class Claude {
    api; // API key provided by the user
    headers; // Headers for Claude requests
    endpoint; // The API endpoint url
    model; // The flag to specify the Claude model
    /**
     * Constructs a new instance of Claude class
     * @param claudeKey: string - The user's Claude authentication key
     * @param model: ClaudeModel - Should be one of the model string provided by Claude
     */
    constructor(claudeKey, model = "claude-3-haiku-20240307") {
        this.api = claudeKey;
        this.endpoint = "https://api.anthropic.com/v1/messages";
        this.model = model;
        this.headers = {
            "x-api-key": this.api,
            "anthropic-version": "2023-06-01", // This is the current version provided by the Claude
            "Content-Type": "application/json",
        };
    }
    /**
     * Generates a response from Claude.
     * @param system: string - The system prompt or context that guides the model's behavior.
     * @param messages: ClaudeMessage[] - An array of ClaudeMessages representing the conversation history.
     * @param temperature (Optional): number - A number between 0 and 1 that controls the randomness of the output. Defaults to 0.
     * @param max_tokens (Optional): number - A number represents the maximum output tokens (words or word fragments). Defaults to 1000.
     *
     * @returns A promise that resolves to a ClaudeResult containing the generated text result.
     * @throws An error if the API request fails.
     */
    async generate(system, messages, temperature = 0, max_tokens = 1000) {
        // Limits the temperature between 0.0 and 1.0
        const validatedTemp = Math.min(1.0, Math.max(0.0, temperature));
        // Limits the minimum max_token to 0
        const validatedToken = Math.max(0, Math.round(max_tokens));
        const data = {
            model: this.model,
            max_tokens: validatedToken,
            temperature: validatedTemp,
            system,
            messages,
        };
        try {
            const response = await axios.post(this.endpoint, data, {
                headers: this.headers,
            });
            return response.data;
        }
        catch (error) {
            if (error instanceof Error)
                throw new Error(error.message);
            else
                throw new Error(String(error));
        }
    }
}