Why ELT Brain exists.

What it actually is

ELT Brain is a small service your AI talks to when you ask it for ELT design or audit help. It doesn't replace your AI – it sits alongside it, supplying the second-language acquisition (SLA) and materials-design expertise that general-purpose AI models don't have on their own.

When you ask Claude, Codex or ChatGPT to design a lesson, write a vocabulary set, generate a reading or listening task, or audit a draft coursebook page, your AI connects to ELT Brain automatically. What comes back is three things:

• The relevant research. The specific findings that bear on the task, with cited sources and an evidence-strength rating on every claim.
• A quality rubric. The review criteria a senior commissioning editor would apply to this kind of work.
• Instructions for your AI to follow. How to structure the output, plus refusal logic for weak goals, mis-pitched material, and unrealistic briefs.

Your AI does the writing. ELT Brain makes sure the writing is structurally sound, level-coherent, and defensible against the rubric a senior editor would apply.

25 tools cover the main jobs of ELT writing and editing: getting the brief right (CEFR level, L1, skill focus, instructional setting), designing single-lesson skeletons and language syllabi, writing vocabulary sets and reading / listening / grammar / speaking tasks, drafting exam items, designing live speaking sessions, supporting classroom activity and learner affect, and auditing draft material. They're organised into six specialist personas: a Brief & Context elicitor and a Learning Scientist, plus four ELT-specific specialists: a Curriculum Architect, an Instructional Writer, a Course Doctor, and a Delivery Coach.

Behind the tools sits a curated knowledge base of 333 research notes: 195 learning-science notes (cognitive load, retrieval, assessment, instructional design, learner psychology), plus 138 ELT-specific notes layered on top spanning 15 domains: SLA foundations (Krashen, Long, Swain, Schmidt, DeKeyser), vocabulary acquisition, grammar instruction, the four skills, language testing, materials and publisher workflows, the CEFR Companion Volume 2020 first-class additions, inclusion and accessibility, teacher development, discourse and pragmatics, and specific contexts (CLIL, EMI, EAL, ESOL, business, EAP, ESP, young learners, exam prep, vocational).

You keep prompting your AI the way you already do, but the quality of what comes back is materially better.

How it pushes back

Most AI tools for ELT produce surface-plausible material. Dialogue that scans. Reading texts pitched roughly to a stated level. Vocabulary sets that look reasonable on a first read. Comprehension exercises with four plausible options each. ELT Brain is designed to resist that in three specific ways.

1. Cited research

Tools cite from the knowledge base, or they refuse. Each note is a complete, cited argument with an evidence-strength rating (strong, moderate, weak, or theoretical). When a tool has nothing substantive to say on a topic, it says so rather than falling back to general AI knowledge. Popular ELT myths – learning styles in language learning, "younger always learn faster," "native speakers are inherently better teachers," the critical period as a hard cliff – get flagged the moment they appear.

2. Specific structural refusals

Some requests don't get built. Content-shaped lesson goals ("cover present perfect") and aspirational ones ("become more fluent") get refused with an explanation and a reframe. Vocabulary sets that mix wildly across CEFR levels get refused. Reading tasks asking five "gist" items on a single text get refused (gist is a whole-text construct – you can ask one or two distinct gist questions, not five). Natural-fast listening tasks at A2 get refused (decoding capacity at the level doesn't support it). Speaking prompts dressed as tasks ("discuss travel") get refused. Each refusal catches a real failure mode the tool surfaced in testing.

3. Senior-editor-grade rubrics

Fifteen ELT-specific rubrics tell the AI what to fix before returning anything. Aspirational lesson goals get sharpened; mis-pitched vocabulary loads get caught; reading items whose form doesn't match their stated sub-skill ("MCQ for gist" where the answer is just a lifted phrase) get flagged; distractor rationales that say nothing ("this might confuse them") get rejected. Every rubric comes with explicit anti-agreeableness instructions: name structural flaws directly, refuse hedge words, keep pushing back when the material is bad.

Rubric compliance is highest on Claude and Codex – that's why those are the recommended AI tools. ChatGPT chat follows the rubrics less consistently on complex audits. The tools still work there, but the best experience is in Claude and Codex.

Why a hand-curated knowledge base

Most AI products that claim to know a domain use a simple retrieval step: take a question, search a pile of documents, hand the model the most relevant chunks. That doesn't work well for ELT, where the primary material is dense applied-linguistics papers, contested methodologies (PPP vs TBLT, explicit vs implicit grammar instruction), and frameworks that evolve over time – the CEFR Companion Volume 2020 added entire scales the 2001 framework didn't have.

ELT Brain uses a different approach. The 138 ELT-specific notes are curated, each as a complete argument: a defensible claim, the primary research behind it, the conditions it holds under, and the design implications that follow. Each note cites primary sources. Each carries an evidence-strength rating. Each links to related notes, so the tools can reason across topics. When a vocabulary question touches lexical coverage, the productive-vs-receptive gap and spaced-recycling notes are one link away.

The knowledge base is publisher-neutral. It draws on Oxford 3000/5000, Cambridge English Profile, NGSL, AWL, ALTE standards, CEFR descriptors, and the wider applied-linguistics literature – without privileging any one publisher's output. Where frameworks compete (which wordlist, which CEFR alignment methodology), the notes name the alternatives rather than picking one.

How it was built and tested

ELT Brain is built on top of a learning science foundation. The 25-tools and 136-note ELT knowledge base were developed for four audiences (freelance writers, commissioning editors, teacher-trainers, in-house publisher editors), with the rubrics calibrated to what a senior commissioning editor catches on a draft coursebook page.

Honest gaps: there's no longitudinal study showing that materials designed or audited with ELT Brain produce measurably better learning outcomes than materials designed without it. The structural differences show up clearly in side-by-side comparisons against published coursebook pages. The direct causal chain from "used this tool" to "learners retained more or acquired faster" takes years and access to publisher production datasets to measure properly. The knowledge base is comprehensive across the principal SLA and ELT-publishing literature; it's thinner on emerging frontiers (large-scale AI in language assessment, multimodal pedagogy beyond text-and-image, very-young-learner content beyond pre-A1 framework guidance). When it's silent on something, the tools say so rather than bluffing.