Rejaviy imtiyozlar

Addon T

Qo'shimcha xususiyat foydalanuvchilari uchun ixtiyoriy reja

Addon T obuna bo'yicha qo'llanma

Obuna bo'yicha qo'llanma

Xususiyat bo'yicha qo'shimcha to'lov solishtirish (-25%)

Review successful usage only. Math similar-question generation applies academy contract prices first; Questi uses actual user plan/generation prices.

{billableCount} billable items / {totalCount} total items

Billing by feature

Billing Invoice

View Book Chat web usage trends by day/week/month/custom range.

Book Chat Trend

View library/Book Chat OCR analysis usage trends by day/week/month/custom range.

Book OCR Analysis Trend

Operator boshqaruv paneli va foydalanuvchi boshqaruvi

Administrator konsoli

Kun/hafta/oy/moslashtirilgan diapazon boʻyicha foydalanish tendentsiyasini koʻring.

Qiyinchilik tendentsiyasi

Ishlab chiqarish/koʻrib chiqish tendentsiyalari va asosiy operatsion koʻrsatkichlarni bir joyda koʻring.

Shu kabi muammoli avlod tendentsiyalarini tahlil qiling

Kontentni tahrirlash

Qo'llash bosilganda o'zgarishlar sahifaga qo'llanadi.

Formulani tahrirlash

Avlod

Mening DB ro‘yxatdan o‘tish tendensiyasini tahlil qiling

Mening DB

Ko'rib chiqishdan foydalanish tendentsiyalarini tahlil qiling

Ko‘rib chiqish

Foydalanuvchilar roʻyxati va funksiya cheklovlarini boshqaring

Foydalanuvchi boshqaruvlari

Persisted activity in the selected period, using the measure shown on each card. Generation, editing and grading requests count separately.

Administrator boshqaruv paneli

Kun/hafta/oy/moslashtirilgan diapazon boʻyicha repetitorning foydalanish tendentsiyasini 1:1 koʻrinishida koʻring.

1:1 Repetitor trendi

Rasmlardan shablon yaratish so‘rovlarining bajarilish natijalarini, shuningdek, haqiqiy API chaqiruvlari, tokenlar va xarajatlarni ko‘ring. Har bir so‘rov ko‘pi bilan 5 sahifani o‘z ichiga oladi.

Rasm→shablon foydalanish statistikasi

Kun/hafta/oy/moslashtirilgan diapazon boʻyicha hujjat konvertatsiyasidan foydalanish tendentsiyasini koʻring.

Hujjatlarni o'zgartirish tendentsiyasi

Review analysis times, requesters, indexing attempts and errors per document. Tokens and costs not reported by File Search remain unavailable.

Book OCR analysis

Review segmentation, generation and modification tokens, calls, agents, costs and errors per solution history.

Handwritten solutions

Historical usage statistics. Past token usage and costs are not backfilled.

User downloads

Kun/hafta/oy/moslashtirilgan diapazonda inshoni tuzatishdan foydalanish tendentsiyasini ko'ring.

Inshoni tuzatish tendentsiyasi

View exam similarity analysis usage trends by day/week/month/custom range.

Exam Similarity Analysis Trend

Kun/hafta/oy/moslashtirilgan diapazonda imtihon tahlilidan foydalanish tendentsiyasini ko'ring.

Imtihonlarni tahlil qilish tendentsiyasi

Review and collect school and mock exam inputs from document conversion, exam analysis, and similarity analysis.

Review and collect exam material

Subject

Exam material collection

Kunlar bo‘yicha natijalar

Har kuni bo‘limlar bo‘yicha muvaffaqiyat va nosozliklarni ko‘ring, nosoz akkauntlar va sabablarni tekshiring, tuzatgandan so‘ng xabarnoma yuboring.

{count} akkauntga xat yuboriladi.

Hal qilingani haqida xabar

Nosozliklar konsoli

Kun/hafta/oy/moslashtirilgan diapazon boʻyicha topilgan savol (manba) foydalanish tendentsiyasini koʻring.

Savol tendentsiyasini toping

Usage and request counts for the selected period.

AI mathematics problem generation

View per-run records and existing usage statistics in one place.

View historical usage and request counts for this subject. Legacy and V2 are different views, not a split by creation date.

View English similar problem generation usage by hour, day, week, month, or a custom period.

English Similar Problems Usage Trend

View Korean similar problem generation usage by hour, day, week, month, or a custom period.

Korean Similar Problems Usage Trend

View math similar problem generation usage by hour, day, week, month, or a custom period.

Math Similar Problems Usage Trend

Inspect each history's actor, LLM usage, cost, and errors from stored operational metadata.

Individual call records load only when details are expanded.

LLM call records

Run details

Errors across all matching runs, ordered by frequency.

Most frequent errors

Runs with stored observability metadata / all runs

Operations overview

Run history

Problem generation runs

Science problem generation

View STEAM problem generation usage trends by day/week/month/custom range.

STEAM Problem Trend

Kun/hafta/oy/moslashtirilgan davr boʻyicha shunga oʻxshash muammo yaratish tendentsiyalarini koʻring.

Avloddan foydalanish

Kun/hafta/oy/moslashtirilgan diapazon boʻyicha qoʻlda yozilgan yechimdan foydalanish tendentsiyasini koʻring.

Qo'lda yozilgan yechim tendentsiyasi

View Labeling API usage trends by day/week/month/custom range.

Labeling API Trend

Kun/hafta/oy/moslashtirilgan diapazonda matematik manipulyativ foydalanish tendentsiyasini ko'ring.

Matematik vositasi trendi

Manbalarni boshqaring, konveyerni ishga tushiring, yangilikni qo‘lda joylang.

Sarlavha va havolani kiriting — AI qisqacha mazmun, muhimlik va turkumni tayyorlaydi. Tekshirib joylang.

Joylanadigan sarlavha

Asl sarlavha

E’lon sarlavhasini o‘zgartirmasdan joylashtiring

Qo‘lda kiritish

Bir marta: yig‘ish → takrorni olib tashlash → AI saralash → rasmiy matn bilan solishtirish → joylash → bildirishnoma.

Yig‘ish konveyeri

Manba

Faqat lentasi bor va robots.txt ruxsat bergan manbalarni qo‘shing.

Manbalar

Ta’lim yangiliklarini boshqarish

Kun/hafta/oy/maxsus oraliq bo‘yicha Mening DB ro‘yxatdan o‘tish tendensiyasini ko‘ring.

Mening DB tendensiyasi

Legacy shows historical exam analysis usage. V2 combines measured tokens, cost, agents, and errors from generation, modifications, and retries for each history. Earlier calls are not backfilled.

Kun/hafta/oy/moslashtirilgan davr bo'yicha foydalanish tendentsiyalarini ko'rib chiqing.

Foydalanishni ko'rib chiqing

Kun/hafta/oy/moslashtirilgan diapazon bo'yicha ob'ektni ko'rib chiqish va yorliqlardan foydalanish tendentsiyasini ko'ring.

Ko‘rib chiqish/yorliq trendi

Combine segmentation, child video generation, and retry tokens, LLM/TTS calls, API cost, and errors per original request. API cost covers metered calls; rendering and storage costs are excluded.

Legacy shows historical video generation requests and completed videos. V2 shows newly measured API usage per history. Earlier calls are not backfilled.

Completed videos and original requests grouped by request time. Child histories do not duplicate the request count.

Video generation

Kun/hafta/oy/moslashtirilgan diapazon boʻyicha sintaktik tahlildan foydalanish tendentsiyasini koʻring.

Sintaktik tahlil tendentsiyasi

Inspect each conversion’s operator, results, tokens, API cost, and errors.

View existing document conversion usage and request counts over time. Legacy and V2 are different views.

Bir yoki bir nechta foydalanuvchi bo‘yicha yaratish/ko‘rib chiqish/Mening DB tendensiyalarini solishtiring.

Foydalanuvchi statistikasi

Kredit o‘zgarishlari

Sahifalangan foydalanuvchilarni koʻrib chiqing, kreditlar bering va har bir foydalanuvchi uchun funksiya cheklovlarini boshqaring.

Maxsus xususiyat qo'shing

Ruxsat etilgan qo'shimcha funksiyalar

Foydalanuvchi kreditlarini ayirish

Foydalanuvchi kreditlarini berish

Foydalanuvchi xususiyatlari cheklovlarini o'zgartirish

Foydalanuvchi a'zoligini o'zgartirish

Obuna tarixi

Barcha foydalanuvchilar

Foydalanuvchilar

Foydalanuvchining foydalanishi haqida xulosa

View lecture video analysis usage trends by day/week/month/custom range.

Lecture Video Analysis Trend

View English vocabulary builder usage trends by day/week/month/custom range.

Vocab Builder Trend

Kun/hafta/oy/moslashtirilgan diapazon boʻyicha taʼlim webtoonidan foydalanish tendentsiyasini koʻring.

Edu Webtoon trendi

Gemini 3.8 Flash plans SVG and GGB corrections; gpt-image-2.5-flare draws the corrected figures. GGB corrections reference the previous SVG correction plan and image. Figures are stored as images, without individually editable points or lines.

Option 2-1 · Gemini + GPT Image 2.5 Flare

Highest difficulty design, independent review and accurate figures · isolated prompts

Advanced: import killer configuration

Killer option

Keep the original image and OCR; skip original SVG reconstruction and cached SVG reuse. Generate and validate the transformed figure with the same option 1-1 settings.

Process argon1015

Previous record: QuestiGeo 1014_1023

Classify the selected crop as geometry, algebraic, physics, or other and cache the result. Classification no longer shows a warning.

Classify before upload

Save each figure crop. Show the figure count multiplied by 500 credits and start processing only after the user agrees to the deduction.

Crop each figure and confirm 500 credits

Convert SVG1 to GeoGebra XML and GGB1, and show GGB1 as soon as it is ready.

SVG1 to GGB1

Load GGB1 and inspect its quality. Create GGB2 and replace the active GGB only when the corrected result is selected; retain GGB1 in history.

Load and review GGB; optionally create GGB2

Store the original image's problem OCR results in a separate JSON artifact.

Save problem OCR separately

Save SVG and GGB candidates with validation history, then settle the credit reservation for completed figures.

Save results and settle credits

Generate SVG1 directly from the crop without a construction plan, then show its thumbnail immediately.

Source crop to SVG1 and thumbnail

Compare the source crop with rendered SVG1. Create SVG2 and replace the active SVG only when a correction is needed; retain SVG1 in history.

Compare source and optionally create SVG2

QuestiGeo Process

Previous record: Process argon1014

QuestiGeo 1014_1023

Reuse the preflight classification cache when present. Without a cache, or when an old other-type result needs checking, classify as geometry, algebraic, physics, or other.

Classify figure type

The direct-upload worker detects the figure region within the chosen crop before OCR. It stops if several figures are found, then normalizes the crop to PNG. A cached detection skips the model call.

Detect and normalize figure crop

Generate GeoGebra XML from the chosen SVG and source image. Production pipeline 091720 selects the SVG candidate before conversion and may call the model again for fidelity repair.

Convert SVG to GeoGebra XML

Static integrity checks and headless GeoGebra loading/capture use no LLM. Visual review of source, SVG, and GGB capture, plus one possible repair, use gemini-3.8-flash. The default gate mode warns and can select the best candidate.

Load, review, and repair GGB

Read problem text, figure flags, and an optional bounding box from the original image. When several figures are present, use the chosen crop for context.

OCR and figure hints from original

Send the cropped image to /figures/classify-sources to classify geometry, algebraic, physics, or other. Pass the result to /figures and save it as the worker classification cache.

Classify the crop before registration

When needed, segment the original page and match the chosen figure to its problem. The selected-crop path for multiple figures can skip this call.

Segment problems and select context

Method A creates an optional matplotlib reference and structure plan. It was enabled in production. Other-type figures skip it; a recitation failure may use the gpt-6.1-sol fallback.

Method A Python reference and plan

Save SVG, GGB, XML, capture, thumbnail, checkpoints, and validation history, then finalize the conversion and credits. Saving calls no LLM.

Save artifacts and settle credits

If another SVG candidate exists, compare both with the source and select the more faithful one. A single candidate needs no comparison call.

Compare SVG candidates

Generate the first SVG candidate from the source figure, problem text, and optional Python reference, then parse, normalize, and check its geometry.

Reconstruct source SVG

When review feedback or the candidate policy calls for another version, generate a new SVG candidate from the original. Cache and validation results can skip this call.

Regenerate an SVG candidate

Compare the rendered SVG with the original for missing objects, intersections, labels, and typography, and provide feedback when another candidate is needed.

Review SVG and regenerate if needed

The browser crops the selected image and uploads both crop and original to storage. Browser processing and file transfer use no LLM.

Select, crop, and upload

Ta'lim yo'lingizni boshqaradigan Questi o'qituvchilari bilan tanishing!
Qaysi o'qituvchi sizga to'liq mos keladi?