Blindspot

Introduction

Blindspot is a browser-based English writing checker that pairs AI-driven corrections with rule-level explanations and follow-up drills. Instead of silently rewriting a sentence, the platform labels the fault, explains the grammar principle behind it, and converts recurring mistakes into targeted practice. Built for learners who know the rules in theory but keep tripping over them in practice, it bundles a text checker, a vocabulary builder and a grammar trainer under a single free account.

Key Features

AI text checking with rule-level feedback. The core checker analyses submitted English and returns a structured result: a correction line where the original phrasing is struck through beside the revised version, plus a “Why it changes” note written in plain language. The landing example corrects “didn't saw” to “didn't see” and then states that “didn't” already marks the past tense, so the following verb stays in its base form. That second field is what separates the tool from a conventional proofreader — the correction is the surface, the rule is the product.

Vocabulary capture with linguistic depth. Suggested words arrive with part-of-speech breakdowns, definitions, example sentences, collocations and synonyms. The demo entry for “endeavour” separates noun and verb senses, supplies collocations such as “make every endeavour” and “creative endeavour”, and lists near-synonyms including attempt, effort, try and strive. Accepted words feed a flashcard queue, and the vocabulary trainer is presented as a first-class section rather than a hidden side panel.

Manual curation of the personal dictionary. Nothing is inserted automatically. The learner decides which suggestions enter the personal dictionary, which keeps the review queue focused on material that matters instead of every token the model flags.

Grammar practice generated from real errors. When the same topic appears repeatedly across checked texts, the system clusters those instances into a named topic and builds an exercise around them. The preview groups three sentences containing the same past-tense fault into a topic called “Past Simple: negative sentences”, adds a one-line rule summary, then asks which verb completes “She didn't ___ to the meeting.” The progression from correction to clustered pattern to quiz is the defining workflow, and it is documented further on the grammar practice page.

Chrome extension for in-context capture. The Chrome extension handles two tasks without leaving the active tab: selecting a word and clicking “+” saves it with a translation, and a writing panel returns corrected text alongside grammar notes. For learners who read and write across many sites, this removes the copy-paste step into a separate application.

Free level assessment. A standalone level test offers a starting benchmark for learners with no clear sense of their current placement, and it is linked directly from the hero section.

Bilingual interface and L1 handling. The product ships with English and Russian locales and recognises Russian words embedded inside English sentences, a practical detail for the primary target audience.

How It Works

The entry point is deliberately small. The homepage form accepts between 30 and 300 characters — enough for a paragraph, an email draft or a chat message — and a live counter reports progress toward the limit. Submitting an under-length or over-length string keeps the “Check my text” button disabled, which avoids wasted round trips.

Triggering the check initiates a sign-in flow when the visitor is anonymous. A window opens for Google authentication, and the draft remains in the browser tab until that process completes. After signing in and confirming a native language, the text is forwarded to OpenAI for analysis and the result opens immediately. Both the submitted text and the returned analysis are persisted to the account, so history stays available for later review. The draft itself is never uploaded before submission, and the privacy policy documents the data handling.

Inside an authenticated session the ceiling rises to 5,000 characters, which comfortably covers long emails, essay paragraphs and short reports. The result view splits into a correction block and an explanation block, with suggested vocabulary presented as optional additions rather than mandatory edits. Grammar topics accumulate in the background and surface in the practice area once a pattern crosses the internal threshold.

The interface follows a single-column, task-first pattern: input at the top, feedback in the middle, and cross-links to the writing, vocabulary and grammar areas beneath each preview. There is no dashboard-heavy onboarding, and a light/dark theme toggle plus a language switch sit in the navigation bar. The whole loop is short — paste, read two fields, decide whether to keep a word, move on.

Use Cases

Intermediate self-learners at B1–B2. The clearest fit. Learners at this stage rarely need a dictionary and rarely benefit from a bare spell-checker; they need to know why “didn't went” is wrong in a way that sticks. The rule-plus-drill loop targets exactly that gap, and the free tier removes the commitment barrier.

Russian-speaking learners. The interface localisation and the recognition of Russian fragments inside English text point to a deliberate focus on L1 interference, where mistakes often stem from direct translation rather than from unknown grammar.

Professionals writing in a second language. Work emails, status notes and client messages usually fit well inside the 5,000-character ceiling. The Chrome extension is the more interesting angle here, since it lets a user check a message in the tool where it is being written instead of copying it elsewhere.

Students and exam candidates. Clustering errors into named grammar topics mirrors the topic-based structure of most exam syllabi, and the multiple-choice format of the drills resembles test prep more than classroom study.

Vocabulary-focused learners. Collocations, part-of-speech splits, synonyms and flashcards make the vocabulary side useful for learners who want productive knowledge rather than passive recognition.

The product is a weaker fit for advanced writers seeking stylistic editing, tone adjustment or long-document proofing, since the tool is optimised for short, frequent, correction-driven sessions.

Pricing & Value

Blindspot is free to start, with no payment card required at registration. The free tier covers text checking and personalised practice, and the landing copy confirms that paying for additional practice volume is an option rather than a prerequisite. The value proposition rests on cost-per-explanation: a single correction costs nothing, and every accepted word or generated drill compounds into a personalised curriculum that a tutor would otherwise have to assemble manually. The main pricing gap is transparency — the public page does not publish tier names, character limits or usage caps for the paid plan, so the practical ceiling of the free account can only be discovered after signup. The advertised metrics on the closing call-to-action also render as placeholders, which suggests the numbers are still being finalised.

Final Verdict

Blindspot takes a defensible position in a crowded correction market. The differentiator is not accuracy — plenty of tools correct “didn't saw” — but the insistence on explaining the rule and then resurfacing that same rule later as practice. That closed loop, correction to clustered pattern to quiz, is what most grammar checkers skip entirely, and it is the reason this tool suits self-directed learners rather than casual proofreading.

The short input ceiling on the public form, the absence of locales beyond English and Russian, and the reliance on OpenAI for processing are the honest trade-offs. None of them break the core use case, but each narrows the audience. For an intermediate learner who writes short texts daily and wants to understand rather than merely fix mistakes, the free tier is an easy trial with a low switching cost. For advanced writers or teams handling long documents, the fit is considerably weaker.