stubborn positions itself as an “AI Referee for Real-World Challenges.” The core idea is that users can “say one line” describing a challenge, and the system turns it into a playable challenge that friends can join, record proof for, and settle. The workflow shown on the page—Say 1 line, Record proof, Settle win—suggests this is not a traditional chatbot, but a lightweight app built around proof and refereeing for real-world tasks.
The captured text shows that stubborn supports multiple model providers. DeepSeek V4 Flash, Kimi K2.5, and Google Gemini Flash appear to be in the free tier or offer free capabilities, while DeepSeek V4 Pro, Kimi K2.6, OpenAI GPT-5.2, Anthropic Claude 4, Gemini Pro, xAI Grok, Mistral, and others are listed as Premium. Typical use cases include push-up videos, plank holds, typing competitions, study streak check-ins, nearby location check-ins, and game scores. Its value lies in structuring challenge descriptions and checking evidence such as video, microphone input, or location check-ins. However, the page does not explain its video recognition, speech recognition, anti-cheating, or dispute/false-judgment handling mechanisms, so output quality still needs real-world testing.
The pricing model is Free + Premium, but specific prices, quotas, storage limits, usage limits, or subscription periods are not disclosed. The page includes “Auto EN 中文,” suggesting that a Chinese interface or language switcher is available. However, there is no further detail on Chinese challenge understanding or adaptation for Chinese-language audio/video scenarios.
The main advantages are its clear product scenario, suitability for lightweight challenges and points-based stakes among friends, multi-model support, and a free entry point, which lowers the trial barrier. The drawbacks are also obvious: it involves video, audio, location, and public challenges, yet there is no visible explanation of privacy and data handling. API support, integrations, customer support, and an appeal mechanism are also missing. In real-world tasks, an AI referee naturally faces risks around camera angles, environmental noise, and cheating.
stubborn is suitable for individual users and small communities that want to run fitness challenges, study check-ins, game score verification, or points-based wagers among friends. The captured text does not make it possible to determine access conditions from China. Since Premium involves models such as OpenAI, Claude, Gemini, and Grok, network access and payment availability may vary by region. Alternatives to consider include Habitica, Strava Challenges, WeChat group/Discord check-in bots, and Feishu or Notion check-in templates.
⚠ This review is compiled from public sources and does not constitute a purchase recommendation. Verify all facts on the vendor's official site. Verify on summoner.world official site.
summoner.world is an Unknown AI Apps provider. TG4G tracks its product information, an overall rating of 5.0/10, and a China-accessibility score of Workable. Click "Visit Official Site" to reach summoner.world directly.