Trending Model:#1Inklingthinkingmachines⬇16kTrending Model:#2Unlimited-OCRbaidu⬇2237kTrending Model:#3Ternary-Bonsai-27B-ggufprism-ml⬇432kTrending Model:#4Laguna-S-2.1poolside⬇3kTrending Model:#5Bonsai-27B-ggufprism-ml⬇1405kTrending Model:#6Qwen3.6-27B-Fable-Fusion-711-Uncensored-Heretic-NM-DAU-NEO-MAX-MTP-GGUFDavidAU⬇63kTrending Model:#7GLM-5.2zai-org⬇545kTrending Model:#8Nanbeige4.2-3BNanbeige⬇0kTrending Model:#9Qwen3.6-35B-A3B-Uncensored-HauhauCS-AggressiveHauhauCS⬇1998kTrending Model:#10krea2-identity-editconradlocke⬇0kTrending Model:#1Inklingthinkingmachines⬇16kTrending Model:#2Unlimited-OCRbaidu⬇2237kTrending Model:#3Ternary-Bonsai-27B-ggufprism-ml⬇432kTrending Model:#4Laguna-S-2.1poolside⬇3kTrending Model:#5Bonsai-27B-ggufprism-ml⬇1405kTrending Model:#6Qwen3.6-27B-Fable-Fusion-711-Uncensored-Heretic-NM-DAU-NEO-MAX-MTP-GGUFDavidAU⬇63kTrending Model:#7GLM-5.2zai-org⬇545kTrending Model:#8Nanbeige4.2-3BNanbeige⬇0kTrending Model:#9Qwen3.6-35B-A3B-Uncensored-HauhauCS-AggressiveHauhauCS⬇1998kTrending Model:#10krea2-identity-editconradlocke⬇0k

Am-I-Openai-Compatible Probe Sniffs Out Broken APIs

By vramkickedin | | 2 min read
Sleek handheld diagnostic probe with a glowing holographic display shows a sharp green checkmark symbol.

Am-I-Openai-Compatible is a new command-line probe that tests any HTTP server claiming OpenAI API compatibility and shows exactly which endpoints work as promised. It runs a fast two-phase audit on every route—first checking if the endpoint exists, then verifying that responses match OpenAI’s official structure. The project also publishes an honest, up-to-date catalog comparing how popular open-source servers like llama.cpp, vLLM, and Ollama actually behave.

Heiervang Technologies built this tool to solve the messy reality that “OpenAI-compatible” means something different on almost every server. Many implementations skip endpoints, return unexpected data, or ignore important details, which can silently break apps and automation. The probe gives server operators and integrators a clear pass/warn/fail report before anything goes wrong in production.

Instant endpoint audits with minimal traffic

Key Features
  • Audits 27 endpoints in under a second.
  • Phase A confirms route existence, not 404.
  • Phase B validates response shapes against spec.
  • Extends to HT-compat features like reranking.
  • Reports results as JSON or formatted tables.
  • Safe probing limits: two requests per endpoint

This inspection utility is for anyone running local AI servers who needs to know whether their front-end tools, chat UIs, or code libraries will actually connect without errors. Developers building proxies or middleware that stitch together multiple engines will also find it valuable for catching discrepancies early. With one command you get a transparent scorecard instead of relying on vague compatibility claims.

Developer notes and current limitations

The probe is not on PyPI yet, so you install it directly from GitHub using pip with a version tag for reliable CI use. By design, it never waits for long-running tasks like video creation—only the job submission itself is tested to keep scans fast. Heiervang Technologies welcomes contributions and is considering optional profiles for Anthropic-compatible and other API signatures in the future.

"The term 'OpenAI-compatible' gets used so loosely it's nearly worthless." — Source: GitHub

Project Links