Does it work offline?

Yes. Draw+ for Windows is local-first by design. Modeling, validation, traceability, simulation, local saving and the built-in rule-based prompt interpreter all work with no network. The network is used only for sign-in and subscription checks, for cloud save if you use it, and for Cloud AI if you opt in. It runs in air-gapped environments.

Where is my design data stored?

By default, on your PC as local XML text files — nothing is uploaded. Because it's text, you can version, diff and branch it directly with Git. If you sign in, you can additionally save projects to your own account and reopen them on another PC. That cloud copy is byte-identical XML, lives only under your account, and is never shared with anyone else.

If I use the AI, is my design sent anywhere?

Not by default. With Cloud AI off, prompts are handled by a deterministic rule-based interpreter compiled into the app — it runs in-process, needs no network and sends nothing out. It is not an AI model, and Draw+ tells you so wherever its output appears. There is no bundled language model to download, which is why the installer is only about 4 MB. Cloud AI is off out of the box; the first time you use it, a consent dialog explains that your prompt and current model would be sent to an external Cloud AI service, and you can turn it back off anytime from the AI menu. Whenever Cloud AI is off, capped, offline or failing, Draw+ falls back to that offline interpreter and labels the result as not AI-generated.

Is there a limit on AI usage?

Cloud AI is capped at 20 requests per day per user, and the cap is enforced on the server rather than only in the app, which keeps usage and cost predictable. The built-in offline interpreter is not capped — when you reach the daily limit, Draw+ continues with it instead of stopping, and marks that output as not AI-generated.

Does Draw+ ever contact the network without me asking?

Once, and only if you are signed in: at launch it fetches a metadata-only list of your own cloud projects so they appear in the start list next to your local ones. It is a list of names and timestamps — no project content is downloaded until you actually open a row, and nothing is ever written without you pressing something. It runs on a background thread so it can't delay startup, and there is a checkbox on the start page to turn it off. If you never sign in, or you switch it off, Draw+ makes no cloud calls at all.

Can I use Draw+ for free?

Draw+ is free to use for as long as you like, with no sign-in and no time limit. What the $10/month subscription unlocks is persistence: saving models (locally or to your own account) and exporting — pictures, model interchange files and report CSVs.

What happens to my cloud projects if I cancel?

You keep access to them. Saving new versions to the cloud requires an active subscription, but listing, opening, downloading and deleting your own projects only requires you to be signed in. A cancelled subscription never locks you out of data you already saved.

Yes. Purchase and cancellation happen through Stripe's hosted Checkout and Customer Portal in your browser. You can cancel yourself anytime; the app never stores your payment details.

Draw+ is built specifically for individual developers. It does not provide team management, co-editing, shared workspaces, organization accounts or RBAC. Cloud save is per-person: your projects live under your own account only. It's a simple, single account for one person.

What languages is the app available in?

Seven: English, 日本語 (Japanese), 简体中文 (Simplified Chinese), Deutsch (German), Français (French), 한국어 (Korean) and Español (Spanish, added in 1.0.8). You pick one on first launch and can change it anytime in Settings; menus, dialogs, palette, shape names and log messages all follow. The installer also picks its language from your Windows display language, falling back to English.

What diagrams can I draw?

SysML V2 across 10 diagram kinds (bdd / ibd / pkg / par / act / sd / stm / uc / req / test — the nine SysML diagram types plus Draw+'s own test diagram) plus UML, BPMN, mind maps and basic shapes, all on the same canvas.

How do I know the SysML support is actually correct?

Draw+ was reviewed against the standard SysML reference text one chapter at a time — package, block definition, internal block, parametric, activity, interaction, state machine, use case and requirement diagrams, plus cross-cutting allocation and domain customization. Each review compared the notation, permitted elements, endpoint rules and documented pitfalls of a diagram kind against the implementation, and the differences were closed. What you get from that is practical rather than academic: the palette only offers elements that belong on the diagram you are editing, and the validator explains, per element, why something doesn't.

Does Draw+ handle allocation and other cross-cutting relationships?

Yes. Allocations are classified across all five kinds — requirement, functional/behavioral, flow, structural and property — and separated by level, so a definition-level allocation is distinguished from a usage-level or asymmetric one. Draw+ also checks the allocations themselves: self-allocations, circular chains, redundant duplicates and nonsensical targets are reported as problems alongside the rest of the diagram's validation results.

Can I customize SysML for my own domain?

Yes, using SysML's own extension mechanism. You can define stereotypes and metaclasses with extensions and required flags, organize them into profiles, apply several stereotypes to a single element, collect reusable elements into model libraries, and define views and viewpoints with stakeholders, concerns, method and presentation. This is the language-level profile mechanism — Draw+ does not sell industry compliance packs.

Can I measure the quality of my model?

Two reports do this from different angles. The dashboard report covers a project's traceability health — coverage percentage, verification progress, relationship counts, plus quality indicators like orphan requirements, duplicate IDs and missing fields. The MBSE assessment scores the whole model instead of one diagram: element and relationship counts, how many of the four modeling pillars actually carry content, and satisfy, verify and allocation coverage. Both are deterministic and computed locally, and every table exports to CSV.

Won't a tool like this be heavy?

No. Draw+ is a Win32 + Direct2D native app — not Electron — so it's light and launches instantly. A C++ core engine keeps large models smooth to edit, and because the default prompt engine is a rule-based interpreter compiled into the app rather than a bundled language model, the installer itself is only about 4 MB.

What is MBSE, and do I need a specialized tool for it?

MBSE (Model-Based Systems Engineering) is an engineering approach that uses formal models — rather than documents alone — as the primary artifact throughout a project lifecycle. A specialized tool matters because it enforces correct SysML notation, maintains live traceability between requirements, models, and tests, and can simulate behavior directly on the model. Draw+ for Windows covers SysML V2 modeling, requirements traceability, simulation, and AI assistance in a single local-first Windows app for $10/month.

How does Draw+ compare to IBM Rhapsody or Cameo Systems Modeler?

Draw+ for Windows and IBM Rhapsody both support SysML V2 modeling and requirements traceability, but differ sharply in cost, deployment, and scope. Rhapsody and Cameo are enterprise suites, licensed per seat per year at enterprise price points and bought through a procurement process — check each vendor's current list price for exact figures; Draw+ is a flat $10/month with no per-seat pricing. Rhapsody requires cloud connectivity or large enterprise infrastructure; Draw+ is local-first and works in air-gapped environments. Draw+ is designed for individual engineers who need affordable local MBSE without a multi-year procurement process.

What is SysML V2 and why does it matter?

SysML v2 is the latest OMG standard for systems engineering modeling. It replaces SysML 1.x with a cleaner, text-first kernel called KerML. Draw+ gives you 10 diagram kinds: the nine familiar SysML diagram types — Block Definition (bdd), Internal Block (ibd), Package (pkg), Parametric (par), Activity (act), Sequence (sd), State Machine (stm), Use Case (uc) and Requirement (req) — plus a Test diagram (test) of its own for test-case models, each with a methodology-aware palette and validation. Underneath, elements follow the SysML v2 Definition–Usage model and export as SysML v2 text (.sysml / .kerml).

Can Draw+ be used for safety-critical engineering such as ISO 26262 or DO-178C?

Draw+ provides requirements-to-test traceability and verification coverage reporting that are directly relevant to safety-critical workflows under ISO 26262, DO-178C, IEC 62304, and similar standards. Draw+ does not ship compliance checklists, standard-specific annex output or evidence packages, and the application is not a certified or qualified tool — using it in a formal certification context requires your organization to perform tool qualification independently. What it does give you is the requirement, model and trace structure those standards expect: Satisfy and Verify links, gap analysis, and Git-versioned model history.

Can I move models between Draw+ and other SysML tools?

Yes. A single Export command (Ctrl+E) writes either pictures — PNG, JPEG, SVG, PDF — or models: OMG XMI 2.1, and SysML v2 text (.sysml / .kerml). You can import .xmi and .sysml files back onto the canvas, and the round trip preserves structure, connectors and diagram layout rather than flattening it into loose shapes.

What file format does Draw+ use, and is it Git-compatible?

Draw+ saves models as plain XML text files in its own format (.xml). The format is human-readable, so you can open model files in a text editor, diff them with standard tools, and commit them directly to Git — version history, branching and merging work exactly like source code, with no proprietary binary format and no cloud sync required. For standards-based interchange, Export (Ctrl+E) writes OMG XMI 2.1 or SysML v2 text (.sysml / .kerml).

Can I use Draw+ in an air-gapped or classified environment?

Yes. Leave Cloud AI off (its default) and skip cloud save, and the application runs entirely without network access — modeling, validation, simulation, rule-based prompt assistance and local file saving all work offline. The subscription status is cached locally, so brief connectivity is needed only for renewal checks. Day-to-day modeling is fully air-gapped compatible.

Does Draw+ support code generation?

Code generation is on the roadmap but is not available in version 1.0.8. The current release focuses on SysML V2 modeling, requirements traceability, three built-in simulators (activity token flow, sequence timelines, parametric constraint solver), and AI assistance. You can export diagrams and models to PNG, JPEG, SVG, PDF, XMI and SysML v2 text for documentation, interchange and client deliverables.

How does Draw+ get updates?

The copy you download from this page checks for updates itself and can install them for you. Every downloaded installer is verified against an EdDSA signature before it is applied, so an update that has been tampered with is rejected rather than run. A copy installed from the Microsoft Store is updated by the Store instead, on the schedule Windows uses for your other Store apps.

Should I install from the Microsoft Store or download the installer?

Both channels ship the same application, so pick whichever fits how you manage software. The Microsoft Store installs and updates Draw+ for you and never shows a SmartScreen prompt. The direct installer is a standalone, code-signed .exe pinned to one exact version — the better fit for offline machines, for a controlled rollout, and for keeping a copy of the artifact you deployed.

KerML (Kernel Modeling Language) is the foundational semantic layer underlying SysML V2. It defines the core constructs — types, features, relationships, and behaviors — that SysML V2 diagrams are built on. Draw+ can show the current diagram as KerML text, and export the model as SysML v2 text (.sysml / .kerml) for tools that read the OMG standard directly.

Is Draw+ available for macOS or Linux?

Draw+ for Windows currently runs on Windows 10 and Windows 11 (64-bit). A separate macOS edition of Draw+ is also available. Linux support is not currently planned.

オフラインで使えますか？

はい。Draw+ for Windows はローカル完結が前提です。モデリング・検証・トレーサビリティ・シミュレーション・ローカル保存・組み込みのルールベースインタプリタは、すべてネットワークなしで動作します。通信が必要なのは、サインインとサブスクリプションの確認、クラウド保存を使う場合、そしてクラウド AI にオプトインした場合だけです。エアギャップ環境でも使えます。

設計データはどこに保存されますか？

既定では、あなたの PC のローカルファイル（XML テキスト）です。アップロードは行われません。テキスト形式なので、Git でそのまま差分管理・履歴追跡できます。サインインすれば、加えてプロジェクトを自分のアカウントに保存し、別の PC で開き直すこともできます。そのクラウド上のコピーはバイト等価の XML で、あなたのアカウント配下にのみ存在し、他者に共有されることはありません。

AI を使うと、設計が外部に送られますか？

既定では送られません。クラウド AI がオフのとき、プロンプトはアプリに組み込まれた決定論的なルールベースのインタプリタが処理します。プロセス内で動作し、ネットワークを必要とせず、何も外部に送りません。これは AI モデルではなく、Draw+ はその出力が現れる場所でその旨を明示します。ダウンロードすべき言語モデルを同梱していないため、インストーラーも約 4 MB に収まっています。クラウド AI は初期状態でオフで、初回利用時には「プロンプトと現在のモデルを外部のクラウド AI サービスへ送信する」旨の同意ダイアログが表示され、AI メニューからいつでもオフに戻せます。クラウド AI がオフのとき・上限到達時・オフライン時・失敗時には、このオフラインのインタプリタにフォールバックし、その結果が AI 生成ではないことを明示します。

AI の利用回数に制限はありますか？

クラウド AI には 1 日あたり 20 回／ユーザーの上限があり、しかもアプリ側だけでなくサーバー側で適用されるため、利用量とコストが予測可能な範囲に収まります。組み込みのオフラインインタプリタに上限はありません。1 日の上限に達しても停止するのではなく、そのインタプリタで処理を続け、その出力が AI 生成ではないことを明示します。

こちらが操作しなくても通信することはありますか？

サインインしている場合に限り、1 回だけあります。起動時に、自分のクラウドプロジェクトのメタデータのみの一覧を取得し、ローカルのプロジェクトと並べて起動リストに表示します。取得するのは名前と更新日時だけで、プロジェクトの中身は実際にその行を開くまでダウンロードされません。書き込みは常に明示的な操作が必要です。処理はバックグラウンドスレッドで走るため起動を遅らせることはなく、スタートページのチェックボックスでオフにできます。サインインしない場合、またはオフにした場合、Draw+ はクラウドへ一切通信しません。

無料で使えますか？

Draw+ は、サインインなしで、期限もなく、好きなだけ無料で使えます。月額 10 USD のサブスクが解放するのは「永続化」です。モデルの保存（ローカルファイルでも、自分のアカウントでも）と、エクスポート（画像・モデル交換ファイル・レポートの CSV）が対象になります。

解約したら、クラウドのプロジェクトはどうなりますか？

引き続きアクセスできます。クラウドへ新しく保存するには有効なサブスクが必要ですが、自分のプロジェクトの一覧・閲覧・読み込み・削除は、サインインしているだけで行えます。解約によって、すでに保存したデータから締め出されることはありません。

解約は簡単ですか？

はい。購入・解約はブラウザで開く Stripe のホスト型ページ（Checkout / カスタマーポータル）で完結します。いつでも自分で解約でき、アプリ内に決済情報を抱え込みません。

チームで使えますか？

Draw+ は個人開発者向けに特化しています。チーム管理・共同編集・共有ワークスペース・組織アカウント・権限管理（RBAC）は提供していません。クラウド保存も個人単位で、プロジェクトは自分のアカウント配下にのみ置かれます。一人ひとりが、自分の手元で完結して使うためのツールです。

アプリは何語に対応していますか？

7 言語です。English・日本語・简体中文（簡体中国語）・Deutsch（ドイツ語）・Français（フランス語）・한국어（韓国語）に、1.0.8 で追加した Español（スペイン語）を加えた 7 言語に対応します。初回起動時に選択でき、設定からいつでも変更できます。メニュー・ダイアログ・パレット・シェイプ名・ログメッセージまで追随します。インストーラーも Windows の表示言語に合わせて表示され、該当しない場合は英語になります。

どんな図が描けますか？

SysML V2 の 10 図種（bdd / ibd / pkg / par / act / sd / stm / uc / req / test — SysML の 9 図種に Draw+ 独自のテスト図を加えたもの）に加え、UML・BPMN・マインドマップ・基本図形に、同じキャンバスで対応します。

SysML への対応が本当に正しいと、どう確認できますか？

Draw+ は SysML の標準リファレンス文献に照らし、章を 1 つずつ突き合わせて検証しています。パッケージ図・ブロック定義図・内部ブロック図・パラメトリック図・アクティビティ図・相互作用（シーケンス）図・ステートマシン図・ユースケース図・要求図に加え、図をまたぐアロケーションとドメインカスタマイズまでが対象です。各回で記法・許容要素・端点規則・文献が挙げる落とし穴を実装と突き合わせ、差分を埋めました。その成果は学術的なものではなく実用的です。パレットは編集中の図に属する要素しか出さず、検証器は要素ごとに「なぜ許されないのか」を説明します。

アロケーションなど、図をまたぐ関係にも対応していますか？

はい。アロケーションは 5 種（要求／機能・振る舞い／フロー／構造／プロパティ）に分類され、さらに定義レベル・使用レベル・非対称のいずれかとして区別されます。加えて Draw+ はアロケーション自体も検査します。自己アロケーション、循環、冗長な重複、意味の通らない宛先は、その図の他の検証結果と並んで問題として報告されます。

SysML を自分のドメイン向けにカスタマイズできますか？

はい。SysML 自身の拡張機構を使います。拡張や required 指定を伴うステレオタイプ／メタクラスを定義し、それらをプロファイルにまとめ、1 つの要素に複数のステレオタイプを適用し、再利用要素をモデルライブラリに集約し、ステークホルダー・関心事・手法・提示形式を伴うビュー／ビューポイントを定義できます。これは言語レベルのプロファイル機構であり、業界別のコンプライアンスパックを販売するものではありません。

モデルの品質を測ることはできますか？

2 種類のレポートが、それぞれ別の角度から測ります。ダッシュボードレポートはプロジェクトのトレーサビリティの健全性を扱い、カバレッジ率・検証の進捗・関係種別ごとの件数に加え、孤立した要求・ID の重複・欠落フィールドといった品質指標を示します。MBSE 評価は 1 枚の図ではなくモデル全体を採点し、要素数と関係数、4 つの柱のうちいくつが実際に内容を持つか、Satisfy／Verify／アロケーションのカバレッジを算出します。どちらも決定論的にローカルで計算され、表はすべて CSV に書き出せます。

動作が重そうですが？

いいえ。Draw+ は Win32 + Direct2D のネイティブアプリで、Electron ではないため軽快で起動も一瞬です。C++ で書かれたコアエンジンが大きめのモデルでもなめらかな編集を支え、既定のプロンプト処理も言語モデルではなくアプリに組み込まれたルールベースのインタプリタなので、インストーラー自体も約 4 MB です。

MBSE とは何ですか？専用ツールは必要ですか？

MBSE（モデルベースシステムズエンジニアリング）は、ドキュメントだけでなく正式なモデルをプロジェクト全体の主要成果物として使用するエンジニアリング手法です。専用ツールが重要なのは、正しい SysML 記法を強制し、要求・モデル・テスト間のトレーサビリティをリアルタイムに維持し、モデル上で直接シミュレーションできるためです。Draw+ for Windows は、SysML V2 モデリング・要求トレーサビリティ・シミュレーション・AI 支援をすべて 1 つのローカル完結 Windows アプリに収め、月額 10 USD で提供します。

IBM Rhapsody や Cameo Systems Modeler と比べてどうですか？

Draw+ for Windows と IBM Rhapsody はどちらも SysML V2 モデリングと要求トレーサビリティをサポートしていますが、コスト・展開方法・対象規模が大きく異なります。Rhapsody や Cameo は、シート単位・年額のエンタープライズ価格で調達プロセスを経て導入するスイートです（正確な金額は各ベンダーの最新の価格表をご確認ください）。一方 Draw+ は per-seat 課金なしの月額 10 USD（フラット）です。Rhapsody はクラウドや大規模インフラを前提としますが、Draw+ はローカル完結でエアギャップ環境でも動作します。手の届く価格でローカル MBSE を必要とする個人エンジニアのためのツールです。

SysML V2 とは何ですか？なぜ重要なのですか？

SysML v2（Systems Modeling Language バージョン 2）は、システムズエンジニアリングモデリングのための最新 OMG 標準です。SysML 1.x を刷新し、KerML というクリーンなテキストファースト基盤を採用しています。Draw+ が備えるのは 10 図種です。おなじみの SysML 9 図種 — ブロック定義図（bdd）・内部ブロック図（ibd）・パッケージ図（pkg）・パラメトリック図（par）・アクティビティ図（act）・シーケンス図（sd）・状態機械図（stm）・ユースケース図（uc）・要求図（req）— に加え、テストケースのモデル化のための Draw+ 独自のテスト図（test）を用意し、いずれも図種連動パレットと検証器を伴います。要素は内部的に SysML v2 の Definition–Usage 構造に従い、SysML v2 テキスト（.sysml／.kerml）として書き出せます。

ISO 26262 や DO-178C などの安全クリティカルな開発に使えますか？

Draw+ は、ISO 26262・DO-178C・IEC 62304 などの安全規格に関連する要求↔テストのトレーサビリティと検証カバレッジレポートを提供します。ただし、規格別のチェックリスト・Annex 出力・Evidence パッケージは搭載しておらず、アプリ自体もこれらの規格における認定・適格化ツールではありません。正式な認証プロセスでの使用には、組織による独自のツール適格化が必要です。提供できるのは、規格が前提とする要求・モデル・追跡の構造です（Satisfy / Verify リンク、ギャップ分析、Git 版管理による設計履歴）。

他の SysML ツールとモデルをやり取りできますか？

はい。1 つのエクスポートコマンド（Ctrl+E）で、図（PNG・JPEG・SVG・PDF）にも、モデル（OMG XMI 2.1、および SysML v2 テキスト .sysml／.kerml）にも書き出せます。.xmi・.sysml はキャンバスに読み戻すことができ、往復しても構造・コネクタ・図のレイアウトが保持されます（バラバラの図形に崩れません）。

Draw+ が使うファイル形式は何ですか？Git と互換性はありますか？

Draw+ はモデルを、独自形式のプレーンな XML テキストファイル（.xml）として保存します。テキスト形式なので、テキストエディタで開いたり、Git で差分を取ったり、ソースコードと同じリポジトリにコミットしたりできます。設計のバージョン管理・ブランチ・マージがコードと同じ流儀で動き、独自バイナリ形式もクラウド同期も不要です。標準に基づくモデル交換が必要な場合は、エクスポート（Ctrl+E）から OMG XMI 2.1 または SysML v2 テキスト（.sysml／.kerml）に書き出せます。

エアギャップ環境や機密性の高い環境で使えますか？

はい。クラウド AI をオフのまま（＝既定のまま）にし、クラウド保存を使わなければ、アプリはネットワークなしで完全に動作します。モデリング・検証・シミュレーション・ルールベースのプロンプト支援・ローカル保存はすべてオフラインで機能します。サブスクリプション状態はローカルにキャッシュされるため、更新確認時のみ短時間の接続が必要ですが、日常のモデリング作業は完全にエアギャップ対応しています。

Draw+ for Windows — Privacy Policy (Handling of Personal Information)

Service: Draw+ (Draw+ for Windows, the "Service")

Established: June 12, 2026　Last revised: August 16, 2026

1. Basic Policy

shokoma (the "Company") recognizes the importance of users' personal information in providing the Service, complies with the Act on the Protection of Personal Information ("APPI") and other applicable laws and guidelines, and handles personal information appropriately in accordance with this Privacy Policy (this "Policy").

The Service is designed around a "local-first" principle: the artifacts you create and edit — diagrams, models, and the like — are stored by default in local files on your own device. Except for the limited information described below (such as account and billing status), and any project you explicitly choose to save to cloud storage under your own account, the Company does not collect or retain your artifacts themselves (Section 10).

2. Business Operator Information (Personal Information Handling Operator)

Operator: shokoma
Address: Disclosed without delay in writing or by email upon request.
Representative: Shohei Komatsu
Personal Information Protection Manager: Shohei Komatsu
Contact: https://shokoma.com/contact

3. Definitions

Personal information: "personal information" as defined in Article 2(1) of the APPI — information about a living individual that can identify the specific individual by name, email address, or other description (including information that can be readily collated with other information, thereby identifying the individual).
User: anyone who uses the Service (whether or not they have created an account).
Account: the user-identifying information tied to an individual user, created by signing up / signing in to the Service.
Artifacts: the diagrams, models, nodes, text, and other data that a user creates or edits using the Service.

4. Information We Collect and How

The Company collects the following information to the extent necessary to provide the Service.

4.1 Account Information (Authentication)

The Service's account feature uses Firebase Authentication (provided by Google LLC). When you sign up / sign in, the following is collected:

Email address
An account identifier assigned by the authentication platform
Password (the Company does not retain plaintext passwords; they are managed securely by the authentication platform)
Authentication-related metadata such as sign-in timestamps and method

4.2 Payment / Subscription Information

Payments for the Service are processed through the payment platform provided by Stripe (Stripe, Inc. and its affiliates).

Payment details such as credit card number, expiration date, and security code are collected and managed directly by Stripe and do not pass through the Company's devices or servers. The Company does not retain this card information.
The Company (or its billing-management service) collects and retains only the minimum information needed to determine billing status: an identifier for your account, an identifier issued by the payment platform, and your subscription state and billing period.

4.3 Usage / Diagnostic Information (Telemetry)

Usage and diagnostic information (telemetry) for quality improvement and defect analysis is not sent by default (opt-in). Only if you explicitly enable it in settings may anonymous or pseudonymized usage statistics and error information be collected. In air-gapped environments, telemetry can be fully disabled.

4.4 Information Associated with AI Features

By default, the Service's AI Co-Pilot commands are handled by a rule-based interpreter built into the application. It runs on your device, is not a generative AI model, and does not send diagrams, models, prompts, or the like externally. The application labels output produced this way as not AI-generated.

Generative AI is provided only through the cloud. Only if you voluntarily select (opt in to) the cloud LLM (a language model provided by an external AI vendor) will the prompts you enter and related model information be sent to that external AI vendor. In that case, handling is also subject to that vendor's privacy policy.
Unless you opt in to the cloud LLM, using AI features does not transmit your artifacts externally.

4.5 Inquiry Information

If you contact us, we collect the content of your inquiry, your email address, the history of correspondence, and similar information.

4.6 Information Stored on Your Device

Artifacts (diagrams, models, etc.), the list of recently opened projects, and locally cached authentication / subscription-status information are stored on your device. These are not collected by the Company, other than a project you explicitly choose to save to cloud storage under your own account (see section 10).

4.7 Reports About AI Output

If you use the built-in feature for reporting AI output you consider inappropriate, the report you submit — including the reported output, any description you add, and the technical context needed to investigate it — is transmitted so it can reach our support contact. You can submit a report whether or not you are signed in.

5. Purposes of Use

To provide the Service, verify identity, and authenticate / manage accounts
To handle subscription sign-up, billing, renewal, cancellation, and status determination (including determining whether the save feature is available)
To charge fees, process payments, and issue receipts / invoices
To respond to user inquiries and provide support
To address defects and to maintain and improve the quality and features of the Service (where telemetry is enabled)
To detect, prevent, and respond to violations of the terms and to misuse
To send important notices (changes to terms / policies, material changes to or failures of the Service, etc.)
To fulfill legal obligations

If the Company handles personal information beyond the above purposes, it will obtain your consent in advance.

6. Provision to Third Parties

The Company will not provide personal information to third parties without your prior consent, except where:

required by law;
necessary to protect a person's life, body, or property and obtaining consent is difficult;
especially necessary to improve public health or promote the sound upbringing of children and obtaining consent is difficult;
necessary to cooperate with a state agency, local government, or their delegate in performing legally prescribed duties, and obtaining consent may impede those duties; or
personal information is provided in connection with a business succession due to a merger or other cause.

7. Subcontractors / External Services

The Company may outsource all or part of the handling of personal information, or integrate with external services, to the extent necessary to achieve the purposes of use, and will exercise necessary and appropriate supervision over such parties. The main external services are:

External service	Provider	Main information handled	Purpose
Firebase Authentication	Google LLC	Email address, user UID, authentication metadata	Account authentication / management
Cloud Firestore (only if you use cloud save)	Google LLC	Contents and metadata (name, timestamps, size) of projects you explicitly save, stored under your own account	Personal-account cloud storage and cross-device access to your own projects
Stripe (payments / Customer Portal)	Stripe, Inc. and others	Payment details (not retained by the Company), customer ID, billing status	Payment / subscription management
Cloud LLM (only if you opt in)	The AI vendor you select	Prompts you enter, and the model you are working on	Providing AI assistance features

Please refer to each provider's privacy policy regarding their handling of personal information.

8. Retention Period

The Company retains personal information only for the period necessary to achieve the purposes of use or as required by law, and promptly deletes or anonymizes it thereafter. If you delete your account, the Company will delete account-linked information it holds within a reasonable period, except information subject to retention obligations under billing, tax, or other laws.

9. Security Management Measures

Delegating authentication and payments to proven external platforms (Firebase Authentication, Stripe) and designing the system so the Company does not retain sensitive information (plaintext passwords, card details)
Encryption in transit (HTTPS) and appropriate protection of stored data
Minimizing access privileges and limiting the personnel who handle data
Keeping authentication / subscription tokens stored locally short-lived and discarding them on sign-out
Limiting data retained by the Company to the minimum needed for billing and authentication (data minimization)

10. Local-First Principle (Artifacts Not Transmitted by Default)

The Service stores your artifacts (diagrams, models, etc.) in local files on your own device by default, and that path is complete on its own. Telemetry is not collected except when you opt in, and signed subscription-status tokens and other data never contain the body of your artifacts.

If you are signed in, you may additionally choose to save a project to cloud storage held under your own account (Cloud Firestore, provided by Google LLC). The following applies to that optional feature:

Uploading is always an explicit action you take on a specific project. There is no background or automatic synchronization of project contents.
The stored data is the same file content as the local file, held under your own account only. It is not shared with other users, is not published, and no share links or web viewer are provided.
Access is enforced on the server so that only the signed-in owner of an account can read, write or delete the projects under it.
Saving requires an active subscription; listing, opening, downloading and deleting your own projects requires only that you be signed in, so cancelling never locks you out of data you have already saved.
You may delete your cloud projects at any time from within the Service.
When you are signed in, the Service performs one metadata-only listing of your own projects at start-up so they appear in the start list. It retrieves names and timestamps only — no project content — and you can switch it off in the Service's settings.

If you do not sign in, or do not use the cloud save feature, no artifact of yours is transmitted.

11. Cookies and Similar Technologies

The desktop application itself does not primarily use cookies in the way a web browser does. However, when a browser is used for payments / Customer Portal (Stripe) or authentication, those external services may use cookies, etc., subject to their respective policies.

12. Your Rights

As provided by law, you may request the following regarding your personal information held by the Company:

Notification of the purpose of use
Disclosure (including provision of electromagnetic records)
Correction, addition, or deletion of content
Suspension of use or erasure
Suspension of provision to third parties

To make such requests, contact us via the channel in Section 16. After verifying your identity, the Company will respond within a reasonable period in accordance with law.

13. Cross-Border Transfer of Personal Information

Some external services used by the Service (Google, Stripe, the cloud LLM vendor you select, etc.) may process information on servers outside Japan. Please understand in advance that such cross-border transfers may occur when you use the Service and opt in to certain features.

14. Use by Minors

If you are a minor, please use the Service (particularly subscribing to the paid plan) only with the consent of a parent or other legal representative.

15. Changes to This Policy

The Company may revise this Policy in response to changes in law or the Service. For material changes, the Company will provide notice by posting on the Service or by other appropriate means. The revised Policy takes effect when posted on the Service.

16. Contact

Operator: shokoma
Attn: Personal Information Protection Manager
Contact: https://shokoma.com/contact