llm.report is an open-source platform for logging and analyzing OpenAI API usage, designed for developers building AI applications. It provides insights into API costs, token consumption, and user-specific expenses, aiming to help optimize prompts and manage budgets.

How It Works

The platform leverages a Next.js frontend with TypeScript and Tailwind CSS, utilizing Shadcn UI for components. It stores data in a PostgreSQL database and handles authentication via NextAuth.js. The architecture is built for self-hosting, with a Docker-based setup for quick local deployment.

Quick Start & Requirements

• Install via cloning the repository and running yarn.
• Requires Docker and Docker Compose for local setup.
• Setup involves copying .env.example to .env, generating NEXTAUTH_SECRET, and running yarn dx.
• Official documentation: Self Hosted Installation, Cloud Installation.

Highlighted Details

• Logs OpenAI API requests/responses for prompt analysis.
• Provides user analytics to calculate cost per user.
• Features a no-code solution for analyzing API costs and token usage.
• Inspired by open-source projects Dub and Plausible.

Maintenance & Community

The project is no longer actively maintained. The founders have moved on due to an inability to find a sustainable business model. Interested parties are encouraged to message the original author on Twitter for potential maintenance or further development.

Licensing & Compatibility

Licensed under the GNU Affero General Public License v3 (AGPLv3) or later. This license may impose copyleft restrictions, requiring derivative works to also be open-sourced under the same license, which could impact commercial or closed-source integrations.

Limitations & Caveats

The project is explicitly stated as no longer actively maintained, posing a significant risk for adoption without a commitment to community-driven support or a new maintainer. The original team was unable to establish a viable business model.