Promptr

Core Functionality: Compresses Large Language Model (LLM) prompts via a single REST API call, saving up to 40% on LLM costs while getting the exact same LLM output.

Main Features:

• Quick Integration: Can be added in just 5 minutes.
• High Compression Rate: The default compression ratio is 0.6 (40% reduction), and it supports customized compression ratios, such as <promptr ratio="0.5"> for 50% compression or <promptr ratio="0.8"> for 20% compression.
• Uncompressed Sections: Supports using the <promptr compress="false"> tag to specify text sections that should not be compressed; uncompressed sections are not billed.

Usage Instructions & API Support: Provides a single API endpoint (http://promptr.me/compress-prompt) and supports quick integration in multiple programming languages:

• curl: Send a POST request with the Prompt in JSON format.
• Python: Make a POST request using the requests library.
• JavaScript: Make a POST request using the fetch API.
• C#: Make an asynchronous POST request using HttpClient. Headers must include accept: text/plain, token: YOUR-API-TOKEN, and Content-Type: application/.

Target Users: Developers, enterprises, and AI application builders who require extensive LLM API calls, especially suitable for scenarios where the same prompt is reused multiple times.

Core Advantages:

• Significant Cost Reduction: When reusing a prompt 100x, using Promptr saves 40% on GPT-4-Turbo and GPT-4. Even when using a prompt only once, it saves 30% on GPT-4-Turbo and 37% on GPT-4.
• Lossless Output: The compressed prompt yields the same LLM output as the uncompressed version.

Pricing Model & Cost: Uses Pay-as-you-go pricing with no hidden costs. Promptr costs $1 per 1M input tokens. For example, using GPT-4 ($30 per 1M input tokens) and reusing a prompt 100x: without Promptr, the cost is $3,000; with Promptr, the LLM cost drops to $1,800 (for 600k input tokens) plus $1 for Promptr, totaling $1,801, which is a 40% saving.