checkfakeapi

CheckFakeAPI provides transparent auditing for large language model (LLM) API relays. Key features include: • Fidelity testing for response accuracy • Latency measurement for performance • Verification of real model architecture • Protocol field and capability behavior checks • Stream consistency and usage metadata analysis This open-source tool allows users to ensure they receive the exact service they pay for from API providers. It helps identify issues such as model downgrades, proxy routing, or altered responses, offering full visibility into your API relays. The platform performs in-depth technical observations for Claude, OpenAI, and Gemini relays, examining aspects like thinking block capabilities, function calling, structured output, and long context support. Each report provides detailed evidence and limitations, ensuring that findings are verifiable and clear. API keys are handled securely, used only for the current task and immediately cleared after detection, with no persistent storage. CheckFakeAPI is ideal for developers, engineers, and organizations who rely on external LLM APIs and need to verify the integrity and performance of their service providers. It empowers users to stop guessing about API quality and start auditing with confidence, promoting trust and transparency in the ecosystem.