René's URL Explorer Experiment


Title: LLM API Check — test any LLM API from your browser: latency, reasoning, tools, JSON, vision

Open Graph Title: LLM API Check — test any LLM API from your browser: latency, reasoning, tools, JSON, vision

X Title: LLM API Check — test any LLM API from your browser: latency, reasoning, tools, JSON, vision

Description: Measure TTFT, tokens/s, cache hits and probe reasoning, tool calling, JSON schema and vision support of OpenAI, DeepSeek, Qwen, Doubao, MiniMax, GLM, Kimi, Gemini and any OpenAI-compatible endpoint. Runs entirely in your browser.

Open Graph Description: Measure TTFT, tokens/s, cache hits and probe reasoning, tool calling, JSON schema and vision support of OpenAI, DeepSeek, Qwen, Doubao, MiniMax, GLM, Kimi, Gemini and any OpenAI-compatible endpoint. Runs entirely in your browser.

X Description: Measure TTFT, tokens/s, cache hits and probe reasoning, tool calling, JSON schema and vision support of OpenAI, DeepSeek, Qwen, Doubao, MiniMax, GLM, Kimi, Gemini and any OpenAI-compatible endpoint. Runs entirely in your browser.

Opengraph URL: https://llmapicheck.dev/

direct link

Domain: llmapicheck.dev


Hey, it has json ld scripts:
{"@context":"https://schema.org","@type":"WebApplication","name":"LLM API Check","alternateName":"LLM API Check","description":"Measure TTFT, tokens/s, cache hits and probe reasoning, tool calling, JSON schema and vision support of OpenAI, DeepSeek, Qwen, Doubao, MiniMax, GLM, Kimi, Gemini and any OpenAI-compatible endpoint. Runs entirely in your browser.","applicationCategory":"DeveloperApplication","operatingSystem":"Any (browser)","browserRequirements":"Requires JavaScript","inLanguage":"en","isAccessibleForFree":true,"offers":{"@type":"Offer","price":"0","priceCurrency":"USD"},"url":"https://llmapicheck.dev/"}
{"@context":"https://schema.org","@type":"FAQPage","mainEntity":[{"@type":"Question","name":"Is it safe to paste my API key here?","acceptedAnswer":{"@type":"Answer","text":"The key is kept in your browser (localStorage, or sessionStorage if you prefer) and is only attached to requests sent to the base URL you configured. The site is static and open source: there is no server that could see your key. Still, use a dedicated key with a spending limit and revoke it when done."}},{"@type":"Question","name":"Why does a provider fail with a network error?","acceptedAnswer":{"@type":"Answer","text":"Browsers enforce CORS. If a provider does not send Access-Control-Allow-Origin headers, the request is blocked before it leaves your machine. OpenAI, DeepSeek, Gemini, Groq, OpenRouter and Mistral allow browser calls; for others see the provider guide, or run a local gateway you control."}},{"@type":"Question","name":"How is TTFT measured?","acceptedAnswer":{"@type":"Answer","text":"From the moment the browser starts the request until the first streamed chunk that carries a token. We report both time-to-first-token (any token, including reasoning) and time-to-first-content-token (what a user would actually see or hear)."}},{"@type":"Question","name":"How do you avoid prompt-cache hits inflating results?","acceptedAnswer":{"@type":"Answer","text":"Every performance request starts with a random nonce in the system prompt and uses a random topic and filler text. The cache test intentionally reuses one prefix so cold vs warm can be compared."}},{"@type":"Question","name":"Which APIs are supported?","acceptedAnswer":{"@type":"Answer","text":"Anything implementing the OpenAI Chat Completions API (POST /chat/completions), including Anthropic's and Gemini's OpenAI-compatible endpoints, DashScope, Ark, and self-hosted vLLM/Ollama."}},{"@type":"Question","name":"Can I test an endpoint that only speaks plain http://?","acceptedAnswer":{"@type":"Answer","text":"Not directly from this HTTPS page — browsers block mixed content, except for http://localhost. Forward the endpoint to localhost with SSH, put it behind HTTPS (Caddy, cloudflared), run this site locally with npm run dev, or allow insecure content for this site in your browser. Chrome can also reach local-network http:// addresses after a permission prompt."}},{"@type":"Question","name":"Does it cost money?","acceptedAnswer":{"@type":"Answer","text":"Only your provider's normal token charges. A full capability run is roughly 40 short requests per model plus two ~2.5k-token cache probes; a performance run depends on the number of runs and max output tokens you choose."}}]}

color-schemelight dark
theme-color#0d0d0d
referrerno-referrer
og:typewebsite
og:site_nameLLM API Check
og:localeen_US
twitter:cardsummary_large_image
og:imagehttps://llmapicheck.dev/og.png
twitter:imagehttps://llmapicheck.dev/og.png

Links:

LLM API CheckLLM API Checkhttps://llmapicheck.dev/
Modelshttps://llmapicheck.dev/models
Performancehttps://llmapicheck.dev/performance
Capabilitieshttps://llmapicheck.dev/capabilities
Message formatshttps://llmapicheck.dev/messages
Resultshttps://llmapicheck.dev/results
Docshttps://llmapicheck.dev/docs
Privacyhttps://llmapicheck.dev/privacy
https://github.com/othorizon/llm-api-check
Add a model https://llmapicheck.dev/models
How it workshttps://llmapicheck.dev/docs
Read the privacy statementhttps://llmapicheck.dev/privacy
Docshttps://llmapicheck.dev/docs
Privacyhttps://llmapicheck.dev/privacy
Source codehttps://github.com/othorizon/llm-api-check

Viewport: width=device-width, initial-scale=1


URLs of crawlers that visited me.