Infera
Infera is a free iPhone learning app built around short, daily lessons that explain how large language models work under the hood. Its stated goal is to move users past prompt-writing tips and toward durable mental models of the machinery between a prompt and an answer, covering tokens, embeddings, attention, transformer blocks, positional encoding, training and loss, fine-tuning and RLHF, inference and sampling, context windows, and the sources of hallucination and other limits. The product is organized as a linear, unlock-based curriculum of 72 lessons, each designed to take a few minutes and to introduce a single concept. Every lesson pairs one plain-language explanation with one diagram, so abstractions such as causal masks, attention heads, and positional encoding are presented visually rather than mathematically. The app explicitly targets learners without a math prerequisite. Each lesson embeds multiple-choice and scenario-based checks. Users must answer before the next lesson unlocks, and correct answers are followed by an explanation of why the answer is right. Progress features include daily streaks, gems earned per completed session, and account-synced progress across devices. The app is published by SNTRVN and distributed through the Apple App Store as a free download for iPhone; the marketing site also links to privacy, terms, and an API health endpoint. Infera positions itself as complementary to AI usage rather than as a prompt cheat sheet, aimed at professionals who already work with AI tools but want to understand why models behave as they do.
- Category
- LLM
- Pricing
- Free
Problem solved
People use AI tools daily but rarely understand the mechanisms behind them, leaving them unable to reason about model behavior, limitations, or failure modes such as hallucination. Infera addresses this gap by delivering one small, self-contained LLM concept per day in plain language with diagrams and comprehension checks, so non-specialists can build accurate mental models without needing a math background or committing to a long course.
Use cases
- Learn how tokenization and vocabulary shape model inputs during short daily sessions
- Build an intuition for attention, causal masks, and multi-head attention without heavy math
- Understand the transformer block and positional encoding as a coherent architecture
- Study how training loss, fine-tuning, and RLHF change model behavior
- Learn why inference settings and sampling affect output variability
- Understand context windows and their practical effect on prompts and long documents
- Explain hallucination and model limits well enough to set realistic expectations at work
- Refresh LLM fundamentals as a professional who already uses AI tools daily
Audience
- Developers
- Product Managers
- Business Users
- Hobbyists
Pros
- Free to download with no stated paywall
- Short lessons designed for a few minutes per day
- Plain-language explanations paired with diagrams, no math prerequisite
- Built-in multiple-choice and scenario checks enforce comprehension before unlocking the next lesson
- Ordered curriculum covers the full path from tokens to hallucination and limits
Cons
- Available only on iPhone via the App Store, with no Android, web, or desktop version indicated
- Gamified streaks and gems may motivate some users but read as filler to others
- Conceptual focus means no hands-on coding or model-building practice
- No visible assessment of lesson accuracy depth beyond in-app quizzes
- Community, instructor support, and content update cadence are not documented on the site
Official website: https://infera-lp.sntrvn.space/
Canonical LambdaBase page: https://www.lambdabase.com/tools/infera