Skip to content
Normaline

For Kazakhstan's science and higher education

From regulation to correct document.

Write the content. Normaline turns funding rules, GOST standards and university regulations into the structure, checks and output around it.

Animation: how a requirement becomes a finished documentWhat the animation shows
  1. The rules arrive as separate documents.
  2. Normaline interprets them once, as an executable package.
  3. Authors write inside the official structure.
  4. Checks prove what a machine can prove, as you write.
  5. One source, three results.
  6. From regulation to correct document.

One application can depend on a ministerial order, the competition documentation, Annex 1, GOST 7.32, a university regulation, and Template_final_2.docx. Every researcher reconciles them alone, and every research office checks the result again.

Asked again for every application

  • Is this the current form?
  • Is the annotation within the limit?
  • Where is the English version?
  • Did you complete Table 4?
  • Are the references numbered correctly?
  • Does every citation point to a source?

How a requirement becomes a document

Interpreted once, by people who know the regime. Used in every document written under it.

  1. Authority

    Orders, competition documentation, standards and university rules: everything a document must satisfy.

    Order 118-нж, Annex 1

  2. Package

    The requirement, interpreted once: parts, fields, official tables, languages, limits and the renderer, each traced to its source.

    authoring.yaml

  3. Authoring

    Authors write inside the official structure. Numbering, captions and cross-references are derived, never typed.

    1. Annotation, KK RU EN

  4. Checks

    Every change is checked against the package. Each finding names its rule and the source that requires it.

    GF.WORD_LIMIT

  5. Outputs

    The package's own renderer builds the authoritative PDF, an editable DOCX and the semantic source.

    PDF, DOCX, source

A template answers one question: where is the latest form?

A package answers the real one: how do I prepare this exact document under this exact regime?

What Normaline handles on a real page

Two pages as their packages render them. Each number marks something the author did not have to do by hand.

Page 12 of an NCSTE grant funding application in English
  1. 1The form's own title, in this version's language.
  2. 2The annotation is required in Kazakh, Russian and English, at most 600 words in each.
  3. 3Fixed fields: the topic is limited to 30 words; dates and amounts are entered once for every language.
  4. 4Items 2.1 to 2.3: 150, 100 and 500 words, and no more than 750 together in each language.
NCSTE grant funding application, English version, page 12
Page 11 of a research report under GOST 7.32-2017, with a chart and a table
  1. 1Cross-references point to the figure, formula, table and appendix themselves; their numbers are derived.
  2. 2The chart is drawn from its data file, not pasted in as a picture.
  3. 3The figure caption, with its label and number, as GOST 7.32 sets them.
  4. 4The table caption above the table, numbered by the template.
Research report under GOST 7.32-2017, page 11

Every check names the rule behind it.

Real checks from today's packages. Where a check comes from the authority it says where; where it is the platform's own, it says that.

CheckWhat it catchesBasis
NCSTE grant funding, 2026-2028
GF.WORD_LIMITAn annotation or explanatory-note item over its word limitRequired by the call, Annex 1
GF.CONCEPT_WORD_LIMITItems 2.1 to 2.3 over 750 words together, in any languageRequired by the call, Annex 1, item 2
GF.LANGUAGE_CHOICENo choice yet between Kazakh and Russian for the main textRequired by the call, §4(2) and §5(1)
GF.FORM_STRUCTUREAn official table whose fixed header was changedForm integrity, Annex 1, Tables 1-9
NCSTE commercialization, 2026-2028
BUDGET.TOTAL_MISMATCHA cost estimate that does not add up to the requested grantArithmetic of the estimate
Research report, GOST 7.32-2017
REPORT.KEYWORDS.COUNTFewer than 5 or more than 15 keywordsRequired by GOST 7.32-2017, 5.3.2.1
Invention patent application
PATENT.TITLE.KAZAKHAn invention title missing in KazakhRequired by the examination rules, order No. 1349, p. 6
PATENT.CLAIMS.SENTENCE_ENDA claim that does not end with a full stopRequired by order No. 1349, p. 24(4)
Every package
REFERENCE.DANGLING_CITATIONA citation that points to no sourcePlatform integrity
TRANSLATION.STALEA translation older than the text it translatesPlatform integrity

One system, five reasons to use it.

Researchers, the institutions around them and the bodies that set the rules use the same package for different jobs.

  • Researchers

    Researchers, PhD students, supervisors and project leads

    “How do I produce the correct document?”

    Open the document and the official structure is already there. Spend your effort on the proposal, not on reproducing the form.

    • The official structure, in place from the start
    • Every language version the call requires
    • Official tables with their fixed headers
    • Numbering, labels and captions, generated
  • Research offices

    Research offices, doctoral schools and grant managers

    “Did everyone use the current form, and is everything complete?”

    Checks prove the deterministic part before a document reaches you, so your review goes where judgment is needed.

    • Required parts are present
    • Word and keyword limits are respected
    • Required tables are complete
    • Every required language version exists
  • Universities

    Universities and research institutes

    “Hundreds of authors, and a different Template_final.docx in every department.”

    Decide which package your institution uses for each regime. Every author then starts from the same executable definition, in your own workspaces.

    • Workspaces for departments, teams and projects
    • Owners, admins, editors and viewers; reviewers comment without editing
    • Packages assigned per workspace, shared or private
    • Your own regimes, such as a dissertation or an internal report
  • Funding bodies

    Competition issuers such as NCSTE

    “Applications arrive with the same avoidable defects.”

    Publish a call together with its executable package. Every university and researcher then prepares against one interpretation and fixes deterministic defects before filing.

    • Missing required content
    • Wrong language combinations
    • Altered official tables
    • Exceeded limits
  • Ministries

    The Ministry of Science and Higher Education and its committees

    “The same rule, interpreted separately by every university.”

    Turn a rule into one verified package and distribute it nationally. Every institution runs the same machine-readable interpretation, in its own workspace, with its own data.

    • One interpretation instead of one per institution
    • Every rule traceable to the order or standard behind it
    • Comparable documents across institutions
    • Institutions keep their own workspaces and data

One application, in every language the call requires.

Kazakh, Russian and English versions live together. Facts are entered once; each language writes its own text.

Language rules of the NCSTE grant funding call, 2026-2028
ҚазақшаРусскийEnglish
AnnotationRequiredRequiredRequired
Explanatory noteKazakh or RussianKazakh or RussianRequired
Requested financingKazakh or RussianKazakh or RussianRequired
Entered once
PeopleAmountsDatesIdentifiers
Written per language
AnnotationNarrativeNames and descriptions

Normaline never machine-translates your text. People record how each version was translated and whether it is current.

First page of the application in Russian
The annotation in Kazakh
First page of the application in English
  • Русский
  • Қазақша
  • English
One combined PDF from the package: the annotation in all three languages, the rest in English and Russian.

Regimes you can write under today.

Each package implements a real document regime. The platform stays the same; only the package changes.

  • NCSTE GF

    Grant funding application, 2026-2028, under order 118-нж

    KK, RU, EN

  • NCSTE

    Commercialization grant application, 2026-2028

    RU, KK

  • ГОСТ 7.32-2017

    Research report

    RU

  • ГОСТ 15.016-2016

    Technical assignment for R&D

    RU

  • Patent Law of the RK

    Invention patent application, under examination rules order No. 1349

    RU

  • СТ РК ГОСТ Р 15.011

    Patent research report

    RU

Universities can publish their own regimes the same way. A doctoral dissertation package is in preparation.

One source. Three results you can use.

Authoritative PDF

Rendered by the package's own LaTeX renderer. This is the document as the regime prints it.

Editable DOCX

For colleagues and institutions that still edit in Word: real tables, equations, lists and footnotes. Pagination may differ from the PDF.

  • Tables
  • Equations
  • Lists
  • Footnotes

Semantic source

Text, tables, references, assets and every language version, with a checksum manifest. No template code inside; your work stays yours.

  • document.json
  • bundle.json
  • variants/
  • assets/
  • manifest.json

Questions institutions ask

Does Normaline replace is.ncste.kz?

No. Normaline prepares the documents. Applications are still filed in NCSTE's own system, which Normaline does not connect to.

Does it judge scientific quality or eligibility?

No. Checks report defects that can be proved from the document and its package. Merit, eligibility and budget justification stay with people.

What happens when the rules change?

The package changes, not every author's copy of a form. Each document records the package version it was written under.

Can our university publish its own regime?

Yes. A package can be private to your workspaces: a dissertation, an internal report or any regime you define. A doctoral dissertation package is in preparation.

Can colleagues still edit in Word?

Yes. Every document exports to an editable DOCX with real tables, equations, lists and footnotes. The PDF stays the authoritative result.

Does Normaline translate text or use AI?

No. Normaline runs no AI model and never translates your text. People write each language and record how each version was translated.

Can we leave, and take our documents with us?

Yes. The semantic source exports text, tables, references, assets and every language version, with no template code inside.

Can we import what we already have?

Yes, from Word and LaTeX, as far as the source allows. The import report lists what came across, what was normalized and what could not be kept.

Write the content. Normaline handles the form.

For researchers, for the institutions that support them, and for the bodies that set the rules.

Accounts are issued by invitation for now. Self-service sign-up opens later.