Frequently Asked Questions

SDK.ac is a free collection of eight browser-based AI utility tools: an AI translator, a text summarizer, a prompt generator, an image prompt builder, a code explainer, an email writer, a hashtag generator, and a writing assistant. Each tool page also includes an in-depth guide explaining the technique behind the tool, so you learn how to get better results from AI models — not just what button to press.

Our tools are structured prompt builders. You describe what you need — the text to translate, the article to summarize, the email situation — and the tool assembles a detailed, instruction-rich prompt that encodes best practices (register, format, constraints, verification steps). You then paste that prompt into the AI assistant you already use, such as ChatGPT, Claude, or Gemini. This design keeps your text on your own device and works with whichever model you prefer.

Paste your source text, choose the source and target languages, and set options such as formality (for example Korean 해요체 vs 합쇼체, Japanese keigo, or the French tu/vous distinction), tone, and terms that must not be translated (brand names, product codes). The tool builds a translation prompt that carries all of that context, which produces far more consistent results than pasting raw text into a chatbot.

It turns a long text into a summarization prompt tuned to your goal: an executive summary, bullet-point key takeaways, a TL;DR of a fixed length, or an abstract in a different language. You control length, audience, and whether the summary should preserve numbers and named entities exactly — a common failure point of naive summarization.

The Prompt Generator walks you through the elements of an effective prompt — role, task, context, constraints, output format, and examples — and assembles them into a clean, reusable prompt template. It is useful both for one-off tasks and for building a personal prompt library you refine over time.

The Image Prompt Builder composes prompts using the vocabulary that text-to-image models respond to: subject, style, medium, lighting, camera and lens terms, composition, aspect ratio, and negative prompts. The output works with Midjourney, DALL-E, Stable Diffusion, and similar generators; you can adjust syntax details (such as parameter flags) for your target service.

Paste a code snippet and the tool builds a prompt asking an AI model to explain it at the depth you choose — line-by-line for learners, or architecture-level for reviewers — including edge cases, potential bugs, and complexity. It is language-agnostic: JavaScript, TypeScript, Python, Java, Go, SQL, and most mainstream languages work well. Never paste proprietary code you are not allowed to share into a third-party AI service.

You specify the situation (request, apology, follow-up, cold outreach, refusal), the relationship between sender and recipient, the desired tone, and the key points that must appear. The tool generates a prompt that produces a complete, culturally appropriate email draft — including proper formality levels for Korean and Japanese business email, which generic chatbot requests often get wrong.

Describe your post and target platform (Instagram, X/Twitter, TikTok, LinkedIn, YouTube), and the tool builds a prompt for generating a balanced hashtag set: a few high-volume discovery tags, mid-size niche tags, and specific long-tail tags, matched to each platform's conventions and recommended tag counts.

The Writing Assistant is a general-purpose companion for drafting, rewriting, expanding, shortening, and proofreading text. You choose the operation, audience, tone, and format, and it constructs the editing prompt. Use it when your task doesn't fit a specialised tool — blog outlines, product descriptions, social posts, or polishing text written in your second language.

Yes. We claim no ownership over prompts or text you create with the tools, and you may use them personally or commercially at your own responsibility. Note that the copyright status of AI-generated content varies by country, similar output may be generated for other users, and the third-party AI service you paste a prompt into may apply its own terms to the generated result. See Article 9 of our Terms of Service.

A well-structured prompt significantly improves output quality, but the final text comes from a statistical language model, and no AI output is guaranteed to be accurate. Treat every result as a draft: read it critically, check facts and numbers, and edit before you use it. Our tool guides describe the typical failure modes of each task so you know what to check.

A hallucination is when an AI model states fabricated information with complete confidence — invented citations, non-existent API functions, wrong dates, made-up statistics. It happens because language models predict plausible text rather than retrieve verified facts. The risk is highest when you ask for specific facts, sources, or figures, and lowest when the model transforms text you provided (translating, summarizing, rewriting). Always verify names, numbers, quotes, and references independently.

No — not without qualified human review. AI translation and summarization are excellent for understanding, drafting, and internal communication, but contracts, court documents, medical information, visa paperwork, safety instructions, and financial filings need a professional. An error in these contexts can have serious consequences, and AI models produce fluent-sounding text even when they are wrong, which makes errors hard to spot.

Large language models are non-deterministic: they sample from a probability distribution over possible next words, so the same prompt can yield different phrasings or structures on each run. This is normal. If you need consistency, make your prompt more specific (exact format, length, terminology), or regenerate a few times and pick the best result.

Key ones: (1) training cutoff — models do not know about events after their training data ends, unless they have web access; (2) bias — models learn from human text and can reproduce social and cultural biases; (3) arithmetic and logic — models can make calculation and reasoning mistakes while sounding confident; (4) context limits — very long inputs may be partially ignored; (5) instruction drift — subtle requirements in long prompts are sometimes missed. Our per-tool guides explain how to mitigate each of these.

Advertising is what allows every tool and guide on SDK.ac to stay free, with no account, no subscription, and no usage limits. We serve ads through Google AdSense and keep placements moderate so they do not interfere with using the tools.

Google, as a third-party vendor, uses cookies (including the DoubleClick cookie) to show ads based on your visits to this and other websites. You can opt out of personalised advertising in Google's Ads Settings at https://www.google.com/settings/ads, and opt out of many other vendors at https://www.aboutads.info/choices. Where required by law, a consent banner lets you refuse advertising cookies before any are set. You will still see ads after opting out, but they will not be tailored to your interests.

No. Tool inputs are processed in your browser and are never shared with advertising systems. Ad cookies relate to page visits, not to the content you type. Blocking ad cookies or using an ad blocker does not break the tools, since they run entirely client-side.

It is processed in your browser, on your device. The tools build their output locally; your input is never transmitted to, logged by, or permanently stored on our servers — there is no server-side database of user text. When you clear the input or leave the page, the text is gone. Full details are in Section 3 of our Privacy Policy.

Not by us. Your text reaches an AI model API only at the moment you yourself copy a generated prompt and paste it into a third-party assistant such as ChatGPT, Claude, or Gemini — and from that point the provider's own privacy policy governs how it is transmitted and retained. If we ever add a tool that calls an AI API directly, we will say so clearly on the tool page before you submit anything, send your text over an encrypted connection solely to generate your result, and never store it permanently.

No. We store no accounts (there is no sign-up), no tool inputs, and no generated outputs. Your text is never used to train, fine-tune, or evaluate any machine-learning model, by us or on our behalf. The only data we handle is aggregated analytics (page views, approximate country, browser type) and standard advertising cookies, as described in our Privacy Policy.

Almost everything lives only in your own browser: tool inputs exist only in page memory, and your theme/language preferences sit in localStorage. Clearing your browser's site data for SDK.ac removes all of it. For anything else — for example deleting an e-mail you sent us — write to [email protected] and we will handle it within the statutory deadlines.

Yes, completely free — no account, no subscription, no paid tier, and no usage caps on our side. The site is funded by advertising. Note that the third-party AI assistant you paste prompts into may have its own free-tier limits (message caps, model restrictions), which are governed by that provider, not by us.

SDK.ac works in all modern browsers — Chrome, Edge, Safari, Firefox, and Chromium-based browsers — on desktop and mobile. The interface is responsive, so the tools are fully usable on phones and tablets. We recommend keeping your browser up to date; very old browsers may not render the tools correctly.

Once a tool page has loaded, the prompt-building logic runs locally in your browser, so brief connection drops usually do not interrupt your work. You do need a connection to load pages initially and, of course, to use any third-party AI assistant with the prompt you generated.

E-mail us at [email protected]. Bug reports are most useful with your browser name and version and the steps to reproduce the problem. We read every suggestion, and several tools on the site started as user requests. We aim to reply within a few business days.