explained.fyi

What It Is

explained.fyi positions itself as “The understanding engine,” with the core goal of bridging the gap between “getting a result” and actually understanding that result. It has two tracks: Learn and Interpret. The former is aimed at learners, turning concepts in math, computer science, science, economics, history, philosophy, and more into interactive explanations. The latter is aimed at AI systems, translating model decisions, risk scores, recommendation results, and similar outputs into understandable, embeddable explanations.

Core Capabilities

For learning use cases, it emphasizes sliders, simulations, step-by-step walkthroughs, and explorable charts. For example, users can adjust R₀ and recovery rates in an SIR epidemic model to intuitively see how the system changes. For AI explanations, it focuses on generating explanations from model outputs, feature values, business context, and the target audience. The site showcases an /interpret API: input model_output, feature_values, context, and audience, and it returns an embeddable explanation with sliders, counterfactuals, and plain-language wording. Potential use cases include credit denial explanations, clinical decision support, internal model audits, and real-time product explanations.

Pricing and Trial

The website currently only states that the product is in private beta and that users can apply for early access. It does not disclose any free tier, plan pricing, enterprise quotes, payment methods, or SLA. At this stage, it feels more like an early product with a clear concept than a mature tool whose procurement cost can be directly evaluated.

Pros and Cons

Its main strength is a clear product direction: instead of stopping at SHAP/LIME-style feature weights, it emphasizes counterfactuals, factor breakdowns, distribution exploration, and natural-language explanations tailored to different audiences. It also introduces a workflow of “AI generation + human expert review,” which should help improve explanation reliability. The limitations are also obvious: it does not explain the underlying models, supported model types, data privacy measures, real customer cases, or evaluation metrics. The actual output quality still needs to be validated through hands-on testing.

Who It’s For and Access from China

It is suitable for learners, course or knowledge-content curators who need to explain complex concepts clearly, as well as finance, healthcare, risk-control, and recommendation-system teams with AI compliance and explainability needs. The site provides no information about access from China, so its availability is unknown; payment methods are also not disclosed. As alternatives, depending on the scenario, users may refer to 3Blue1Brown, Khan Academy, Wikipedia, or technical explainability tools such as SHAP and LIME, though these alternatives do not fully match explained.fyi’s positioning around interactive explanations.