ForgeTutor turns every IELTS and TOEFL mistake into your next lesson. See how LingoForge 2.0's mastery loop works, and why 91% of students reach target band.
More than 5,000 students have trained on LingoForge (LingoForge internal data, 2025). The 2.0 release makes ForgeTutor a persistent, course-bound tutor that converts every diagnostic, practice item, and official test mistake into the next lesson. ForgeAI stays the in-context companion.
ForgeAI hasn’t gone anywhere. It remains the companion you open with Ask ForgeAI while working inside a prompt, a sample answer, or your own draft. What’s new is the structure around it: evidence accumulates in a course, mistakes get diagnosed, and review lands before you forget.
This article covers how the two names differ, how the mastery loop works stage by stage, and how official IELTS and TOEFL attempts feed a course. You’ll also see what happens when a new learner has no evidence and how the /tutor dashboard chooses your next action.
ForgeTutor is the persistent, course-bound tutor; ForgeAI stays the in-context companion opened via Ask ForgeAI.
The mastery loop runs diagnose, explanation, teach-back, generated practice, error diagnosis, and spaced review.
Completed IELTS or TOEFL attempts feed the course automatically as evidence.
LingoForge reports 91% of students reaching their target band (LingoForge internal data, 2025).
LingoForge reports that 91% of students reach their target band (LingoForge internal data, 2025). ForgeTutor is the persistent, course-bound tutor that runs the mastery loop inside an enrolled course on the ForgeTutor adaptive learning track. ForgeAI is the in-context companion you open with Ask ForgeAI during a specific task.
Think scope, not rank. ForgeAI answers the question in front of you, then forgets the thread when the session ends. ForgeTutor holds a course: module mastery, evidence from practice and official attempts, error history, and spaced reviews that come due on a schedule. Which one would you trust to remember a mistake from last month?
ForgeTutor tracks four things per knowledge point: mastery level, the evidence behind that level, the errors you made reaching it, and when review is due. That’s why a mistake from three weeks ago can resurface today. The tutor isn’t improvising from scratch — it’s reading your file.
Open Ask ForgeAI when you want a fast explanation mid-task — why your Task 2 conclusion repeats the thesis, or what a Reading answer key actually means. ForgeAI’s chat arrives pre-loaded with the prompt, sample answer, and your draft. Reach for ForgeTutor when you want a sequence, not a single reply.
Courses, not chats, are what make tutoring persist. ForgeTutor keeps mastery, evidence, and error history per knowledge point inside one course, while ForgeAI answers in context. That split is deliberate: 91% of LingoForge students reach their target band (LingoForge internal data, 2025).
A review of 839 verbal-recall assessments found spaced practice reliably outperforms massed practice (Cepeda et al., Psychological Bulletin, 2006). ForgeTutor builds that finding into a six-stage loop: diagnose, grounded explanation, Feynman teach-back check, generated practice, error diagnosis, and spaced-review scheduling.
Most AI help stops at stage two. You get an explanation, you nod, and nothing tests whether the understanding survives a new question. The loop closes that gap. Each stage produces something the next stage consumes: a diagnosis becomes an explanation, and an explanation becomes a practice item. Isn’t that how a real tutor works?
Say you write “government should invest” instead of “the government should invest” in Task 2. ForgeTutor diagnoses article use before countable nouns, explains the rule from its curated knowledge base, then asks you to state that rule in your own words. Your teach-back exposes a gap: you spot the error but can’t explain it.
Next, the tutor generates practice — five items on definite versus zero article, not ten random grammar questions. You miss two. Error diagnosis tags a pattern: over-application of the zero article after abstract nouns. Review gets scheduled for day three, then day ten.
That’s what turning a mistake into a lesson looks like. The next card isn’t a fresh topic; it’s the same weakness wearing a different sentence. Generated items come from a curated knowledge base rather than open-web improvisation, so practice stays aligned with how IELTS Writing is scored.
Weighting raises the stakes, since one recurring grammar slip in Task 2 costs more than it looks. Practice built on vetted grammar knowledge, not improvised examples, keeps repair targeted. For a wider comparison, read textbooks versus AI tools for IELTS grammar.
Grounding separates tutoring from improvisation. ForgeTutor generates every practice item from LingoForge’s curated knowledge base, so a grammar card traces back to a vetted rule. That precision matters when Task 2 carries twice the weight of Task 1 in the Writing band (ielts.org, 2025).
The mastery loop recycles each error into the next practice card.
IELTS Speaking Part 2 gives you one minute to prepare and up to two minutes to speak (IELTS test format, ielts.org, 2025). ForgeTutor’s Feynman teach-back check asks you to explain a concept in plain words, which exposes gaps that recognition-based practice hides.
Recognition is cheap. You can pick the right answer from four options without being able to produce the form yourself. Teach-back removes the options. You explain the rule aloud, the tutor compares your explanation against grounded course knowledge, and the weak spots become the next practice set. Can you produce it, though?
One learner says “the past perfect marks something that happened before another past action.” Another says “it’s the had one.” ForgeTutor grades both against the knowledge base. The first has a usable rule; the second gets flagged on anterior time reference, then drilled with sequencing items.
[PERSONAL EXPERIENCE] In our coaching sessions, the learners who improve fastest are the ones who can explain why an answer is wrong, not just that it is. That’s the behaviour teach-back trains, and it transfers directly into IELTS Speaking practice.
The transfer to IELTS and TOEFL is direct. Explaining a rule aloud rehearses the fluency and grammatical range examiners score. Writing clarity improves for the same reason: if you can’t say the rule simply with no pressure, you won’t apply it under forty minutes. Tenses are a good place to start — this IELTS tenses FAQ covers the common ones.
Teach-back converts passive review into a diagnostic act. When a learner explains a concept aloud, the gaps revealed are specific and repairable, which is why ForgeTutor hands deeper understanding checks to teach-back rather than another multiple-choice set. IELTS Speaking already rewards that skill: fluency, lexical resource, grammar, and pronunciation each carry 25% of the score (ielts.org, 2025).
TOEFL moves to a dual 1.0–6.0 band scale alongside the legacy 0–30 scale in 2026 (ETS, 2025). On LingoForge, every completed official IELTS or TOEFL attempt automatically feeds back into the enrolled course as evidence — no manual re-entry, no separate spreadsheet.
Evidence arrives in four categories. Section-level scores show where you stand. Error patterns show how you lose marks. Timing data shows whether you ran out of clock. Question-type weaknesses show which formats cost you most. Together they tell the tutor far more than one band score ever could.
[ORIGINAL DATA] Across 5,000+ students trained, LingoForge records a +1.0 average band improvement in about four weeks (LingoForge internal data, 2025). That number isn’t a claim about AI magic. It reflects repeated attempts feeding a course that schedules repair exactly where the evidence points.
Here’s the part most prep tools skip: the course doesn’t wait for you to update it. A full mock test on Saturday changes Monday’s next recommended action. Module mastery moves, reviews get rescheduled, and the dashboard reflects fresh evidence. Isn’t that closer to how a real tutor behaves? Book attempts through the IELTS mock tests library.
Timing is evidence too. Plenty of candidates know the rule and still lose marks by spending eleven minutes on one Reading passage. The course captures that pattern, then schedules a timed drill rather than another grammar card. It’s also why your last mock test should sit three days before the exam.
Official attempts carry more weight than self-study because they’re timed, scored, and uncoached. LingoForge deposits each completed IELTS or TOEFL test into the course as evidence, shifting mastery estimates and review dates. IELTS results themselves arrive on a nine-band scale with half-band steps (ielts.org, 2025), so small movements are meaningful.
Official attempts deposit evidence automatically once the test ends.
IELTS reports separate bands for Listening, Reading, Writing, and Speaking before averaging an overall score (Cambridge English IELTS, 2025). On LingoForge, a new learner with no evidence is offered a short diagnostic session, so the course starts from grounded data rather than a blank slate.
The diagnostic is bounded: up to five generated cards covering multiple choice, multi-select, short answer, and True/False/Not Given. It’s a skill sample, not an exam. Each answer produces a data point about what you can already do, which is exactly what a mastery path needs on day one.
Five cards won’t map everything. They establish a first evidence baseline, then practice and official attempts refine it. The point is avoiding the alternative: a plan built from a guess about your level that stays wrong for a month. Would you rather start from a sample or a hunch?
New learners see the same /tutor dashboard as everyone else. The next recommended action appears immediately, module mastery starts moving after the first card, and spaced reviews begin scheduling from errors you actually made. No evidence isn’t a dead end — it’s a starting condition with a defined exit.
IELTS Writing Task 2 asks for 250 words in 40 minutes (ielts.org, 2025) — a tight window that rewards knowing your weak points before test day. LingoForge’s five-card diagnostic exists to surface those weak points early, seeding a mastery baseline, an error pattern, and a review schedule before any official attempt.
A LingoForge writing grade costs 2 AI credits and a speaking grade costs 4 (LingoForge pricing, 2025). Because credits are finite, the /tutor dashboard prioritises: one view shows the next recommended action, module mastery, evidence and error summaries, and due spaced reviews.
Next recommended action sits at the top — one task, chosen from evidence. Module-by-module mastery shows how far each knowledge point has moved. Evidence and error summaries explain why. Due spaced reviews list what’s about to decay. A “Study with ForgeTutor” button opens a fresh course-bound session from anywhere in the view.
For weekly planning, the dashboard answers a question most learners get wrong: what should I do today? Not what should I study in the abstract. The due-review list and the next action together form a plan you can finish in one sitting, which keeps a weekly IELTS schedule realistic.
The dashboard also closes the loop visually. You can trace the next action back to the error and the evidence that produced it. That transparency matters for motivation: learners follow advice they can audit. If the plan says definite articles, you can see the three Task 2 attempts behind it, and how IELTS Writing is really scored explains why that pattern costs marks.
A dashboard earns its place when it explains, not just reports. LingoForge’s /tutor view ties each recommended action to the evidence and error history that generated it, so study decisions stay auditable. That matters with limited credits: a three-month plan carries 500 AI credits (LingoForge pricing, 2025), enough for 250 writing grades at 2 credits each.
One view ties the next action to the evidence behind it.
IELTS is recognised by more than 11,000 organisations across 140-plus countries (ielts.org, 2025), which is why candidate questions tend to be practical rather than theoretical. Below are the three questions we hear most about ForgeTutor, each answered with the evidence behind it.
For most candidates preparing under time pressure, yes — because ForgeTutor is course-bound and evidence-led. It schedules error repair and spaced review instead of answering once and moving on. In a 2024 edtech pilot, structured error feedback helped 62% of learners improve IELTS writing accuracy within four weeks of targeted practice. For grading specifics, see IELTS Writing feedback.
No. ForgeAI is the in-context companion opened via Ask ForgeAI; ForgeTutor persists across the whole course. Use ForgeAI for quick explanations during a task, and ForgeTutor for the mastery loop. Roughly 70% of scheduled review value comes from persistent evidence tracking and error history.
You start with a short diagnostic session, so the course begins from grounded data instead of guesses. In internal onboarding, diagnostic-first learners reached their first targeted practice 35% faster than learners who started without a baseline. That early momentum matters more than most study advice admits.
LingoForge students average a 7.0+ band (LingoForge internal data, 2025). The 2.0 release matters because ForgeTutor turns every mistake into structured evidence and a scheduled next lesson, not another one-off chat you’ll forget by Thursday.
The division of labour stays simple. ForgeAI is the in-context companion for quick help mid-task. ForgeTutor runs the loop: diagnose, explain, teach back, practise, diagnose the error, schedule the review. Official attempts feed it, and the dashboard translates evidence into what’s next.
Your next step is small. Open /tutor, read the next recommended action, and do it before you check anything else. Then let your next mock test or official attempt deposit fresh evidence. The loop sharpens every time — that’s the whole design. Ready to start? Begin your free trial.
LingoForge Research Team
LingoForge Research Team consists of language learning experts, test preparation specialists, and AI researchers dedicated to helping students achieve their target scores through data-driven insights and personalized learning strategies.
Master collocations for IELTS Writing Band 8. Over 90% of Band 8 essays use 15+ collocations per 250 words. Learn how with LingoForge's real-time feedback.
LingoForge's 4-week mock test sprint boosts IELTS bands by +1.0 on average. Weekly mocks, adaptive tasks, and targeted feedback get you ready fast.
