Price, uptime and model-authenticity comparison for LLM API relay providers
Developers who route OpenAI/Claude traffic through third-party LLM API relay resellers have no neutral way to compare price, model coverage or reliability, so they ask strangers in forums for recommendations. A monitoring-and-directory service that independently probes relays and publishes price/uptime/model-authenticity data would replace that guesswork.
The problem
Developers buying LLM access through relay resellers pick a provider by word of mouth because there is no comparable data; relays go down, change prices, silently swap expensive models for cheap ones, or vanish with prepaid credits, so they waste hours re-testing and re-subscribing and lose money on dead balances.
What could be built
A relay directory plus automated monitoring service: scheduled probes of each provider's latency, uptime, listed models and per-token pricing, a model-authenticity check that verifies a claimed GPT-4/Claude endpoint actually behaves like it, community-reported incidents, and alerts when a subscribed relay degrades so users can switch before they lose work.
Who it's for
Who is talking（1）
Related topics
Summaries are AI-generated. The original words are in the threads above.