GPT_API_free - Utilize GPT models with improved privacy and direct domestic access

Introducing the GPT-API-Free Project

Overview

The GPT-API-Free project is aimed at providing users with free and direct access to a range of OpenAI models including GPT-4, GPT-3.5-Turbo, and others like DALL·E and Whisper, without the need for proxies or specialized network configurations. It offers both free and paid API keys, catering to various user needs and ensuring optimal access and speed, especially for users in domestic environments.

Privacy Commitment

Privacy is a top priority for the GPT-API-Free project. The system does not collect, record, or store any user inputs or outputs from OpenAI servers. The project ensures that no identifying information, such as IP addresses or user-agent strings, is shared with OpenAI or any third parties. However, users should be aware that OpenAI itself retains data for a period of 30 days, as per its data usage policies.

Latest Updates

June 14, 2023: Support for GPT-3.5-Turbo-16K was introduced, and the pricing for paid keys was reduced.
June 15, 2023: Functionality for the latest version functions was added.
June 18, 2023: Whisper model support for speech-to-text was introduced.
August 4, 2023: Free Key support for GPT-3.5-Turbo-16K was discontinued.
November 8, 2023: Added support for TTS models with the 1106 version.
November 19, 2023: Released the gpt-4-1106-preview model at a reduced cost.
November 29, 2023: Free API access for GPT-4 was opened, limited to three uses per day.
April 10, 2024: Support for the gpt-4-turbo model was added.
May 14, 2024: The gpt-4o model gained support.
July 19, 2024: Introduced the gpt-4o-mini model with competitive pricing.
August 7, 2024: Supported the gpt-4o-2024-08-06 model with reduced prices.
October 25, 2024: Added support for the gpt-4o-audio-preview model.

Key Features

Supports a variety of models including GPT-3.5-Turbo, GPT-4, and others for Text Embedding, DALL·E, and Whisper.
The free version provides up to three daily uses of GPT-4 and shares up to 200 uses per day between GPT-3.5-Turbo and GPT-4o-mini.
Offers seamless integration with known software and plugins, compatible with official API standards.
Enables streaming responses for an enhanced user experience.
Dynamic acceleration for local connections eliminates the need for VPNs, optimizing access for users within certain regions.
Free for personal use without any associated costs.

Important Considerations

The free API key is for non-commercial personal, educational, and non-profit research purposes only.
Any commercial use or large-scale model training with a free key is prohibited.
The project regularly bans keys detected as abused. If wrongly banned, users are encouraged to contact the team via the provided communication channels.
There are usage limits on free keys of 200 requests per day per IP or key, client-side managed within shared constraints.

Free API Access

Users interested in free API access must register using their GitHub account to obtain personalized keys. For faster and more stable access to GPT-4, transitioning to paid services is recommended. API requests are relayed through specific host URLs to optimize performance based on user location.

Paid API Access

To ensure project sustainability, a paid API key system is in place. This supports higher stability and availability for power users needing more extensive access. The pricing for paid access is designed to be affordable and enables uninterrupted usage of advanced OpenAI models, particularly GPT-4.

Using the API

Interested users can acquire a free API key by registering through the provided OAuth link or opt for a paid key. The API base URL needs to be set appropriately in the application configurations to direct requests to the GPT-API-Free service. Users can customize plugins and scripts by altering endpoint URLs as shown in the usage examples.

Compatibility with Common Software/Plugins

The API is highly compatible with numerous applications and plugins including:

Python OpenAI Library: Modify API settings in scripts or environment variables.
OpenSource Software like gpt_academic: Adjust the API_URL_REDIRECT for service redirection.
Desktop Apps like BotGem: Update settings within app configurations.
Jetbrains and Zotero Plugins: Follow specific configuration instructions for seamless integration.

The project continues to evolve, integrating new models and functionalities while maintaining commitment to offering accessible and reliable AI services to all users. Participants are encouraged to engage with the community through Issue reports and feedback for continuous improvement.