Product Overview
Spamurai is a spam text detection model launched by TUNiB (튜닙, official website tunib.ai). TUNiB is a Korean startup that creates new connections and innovation with AI technology, committed to building a smarter and more responsible AI ecosystem.
Main Company Directions
According to public information on the official website, TUNiB's business includes:
- The next-generation social network "디어메이트": positioned as an SNS where AI and humans coexist.
- AI ethics guardrail solutions: safety protection for the era of generative AI.
- Polyglot Foundation sLLM: a commercially usable, fully open-source Korean small-scale language model.
- In-house sLLM / multi-persona chatbot: a single model that supports multi-persona dialogue.
- Conversational AI: the official slogan is "WE CREATE CONVERSATIONAL A.I. THAT PEOPLE EMOTIONALLY ENGAGE WITH".
Related NLP Capabilities
The official website lists that TUNiB provides the following NLP API capabilities to protect platforms from AI-generated hate speech and privacy leaks:
- Safety Check: detects 12 categories of harmful content and outputs 4 severity levels and probabilities.
- De-identification: detects and masks 14 categories of personal information.
- Other analysis capabilities: text analysis, image analysis, news analysis, video analysis, emotion prediction, word graph, dialect translation, N-Verse text-to-speech, political tendency prediction.
Spamurai belongs to the text detection category among these, but the official website does not provide specific functional descriptions for this model.
Usage and Applicable Scenarios
Official instructions for Spamurai's installation commands, invocation methods, model interfaces, or applicable scenarios are not yet provided. More information can be obtained through the consultation form on the official website page or by contacting tunibridge@tunib.ai.
Limitations and Missing Information
The current materials do not provide Spamurai's charging standards, pricing, technical architecture, supported languages, deployment platforms, performance metrics, or usage steps. None of the above content is currently provided.
