AS A RULE This project does not use any fallback logic nor does it have any mock responses or placeholer content. Also, we never use streming repsonses from any LLM.

FilePromptForge — GPT Processor

Must Read: https://platform.openai.com/docs/guides/tools-web-search?api-mode=responses

Overview

FilePromptForge is a small toolkit for batch-processing text files with an LLM. It includes:

• A GUI installer (Tkinter) to create install folders, write defaults, copy the main executable, and run a quick test.
• A main processing script that combines prompt files, reads input files, sends prompts to an LLM provider (OpenAI or OpenRouter), and writes responses to an output directory.
• Documentation and helper files for configuration and usage.

Features

• Prompt management: combine multiple prompt files into a single system prompt.
• File processing: process multiple input files sequentially and write model responses to the output directory.
• Configurable provider: supports OpenAI and OpenRouter via configuration or environment variables.
• Logging: configurable logging to console and to timestamped log files.
• Installer GUI: creates required directories, default prompts and config, and can copy the main executable into the install directory.

https://platform.openai.com/docs/guides/tools-web-search?api-mode=responses ###Installation

install dependancies

Usage

1. Prepare directories (the installer GUI can create these):

   • prompts/ (place one or more prompt files, e.g. standard_prompt.txt)
   • input/ (place files you want processed)
   • output/ (responses will be written here)

2. Run the main processor: python gpt_processor_main.py [options]

CLI options supported:

• --config Path to configuration YAML file
• --prompt <files...> One or more prompt filenames (from prompts/ directory)
• --input_dir Directory containing input files
• --output_dir Directory where responses will be saved
• --model Model id to use (overrides config)
• --temperature Temperature for the model
• --max_tokens Max tokens for completions
• --verbose Enable verbose (debug) logging
• --log_file Path to a specific log file

Behavioral details

• If no config file is provided, the script looks for default_config.yaml next to the script and then falls back to environment variables (.env) and sensible defaults.
• If no prompt files are passed on the CLI, the script loads all files in the prompts/ directory.
• If no API key is configured, the script will generate mock responses rather than calling the remote API.
• The script processes input files sequentially and writes each response to response_<original_filename> in the output directory.