Frequently Asked Questions

SynthQuery’s AI Content Detector examines your text at multiple levels. It combines statistical signals—such as how predictable word sequences are (often discussed as perplexity) and how uniform or varied phrasing appears (sometimes called burstiness)—with classifiers trained to distinguish patterns common in machine-generated versus human writing. You receive an overall label such as Human, AI, or Mixed, plus per-sentence scores and a heatmap so you can see which lines drive the assessment. Independent benchmarking studies of AI-text detectors, including work by Weber-Wulff and colleagues, show that no tool is infallible: edited AI text, short passages, or unusual style can mislead classifiers, so results should support judgment rather than replace policy decisions. DeepScan runs a heavier analysis pass for mixed or difficult cases and is aimed at subscribers who need extra scrutiny. Your input is processed for the check and is not used to train third-party foundation models through this workflow.

The Humanizer rewrites AI- or template-heavy text so it reads more like natural human prose while preserving the underlying meaning and factual claims you intend to keep. It targets repetitive transitions, generic hedging, and overly uniform sentence rhythm—patterns that both readers and statistical detectors often associate with large language model output. You can pair it with modes suited to academic, professional, or creative contexts and, on many workflows, compare original and rewritten versions side by side before publishing. For illustration, a draft that opens with “In today’s fast-paced digital landscape…” might be steered toward concrete specifics and varied sentence openings without changing the core argument. OECD Programme for International Student Assessment (PISA) data consistently tie reading engagement to clarity and relevance, which is why many editors push AI-assisted drafts toward more concrete, audience-specific language rather than stock phrasing. After humanizing, you can run the same text through the AI detector to see how signals shift, remembering that no detector is perfect. In short, the Humanizer is an editing assistant for tone and naturalness, not a substitute for fact-checking or citation.

SynthQuery’s plagiarism checker measures how much of your text overlaps indexed online sources and, on supported plans, documents you upload as a private corpus—similarity is shown as a percentage and as highlighted passages with links or snippets where available. Accuracy depends on what is in the index, how much paraphrasing occurred, and whether ideas were reused without verbatim overlap; the tool is strongest at catching copy-paste and close paraphrase against sources it can see. The International Center for Academic Integrity notes that contract cheating and misuse of AI are evolving concerns in education, which makes both similarity checks and clear institutional policies important. For teams, adding proprietary documents improves recall for internal reuse. Enterprise options can include larger corpora and tailored retention settings. No plagiarism system proves intent by itself: editors and instructors still interpret matches, add citations, or rewrite. Overall, treat the similarity score as a precise overlap signal you combine with context, not as a standalone verdict on misconduct.

SynthRead is SynthQuery’s readability and audience-fit analyzer. It applies many classic and modern metrics—such as Flesch Reading Ease and grade-level estimates, Gunning Fog, SMOG, and related formulas—so you can see whether a draft matches the reading level you want. Rudolf Flesch’s original work (1948) tied syllable and sentence measures to comprehension difficulty; those ideas still underpin plain-language programs in government and health communication. The U.S. Centers for Disease Control and Prevention, for example, routinely recommends short sentences and common words for public health materials so more people can act on guidance. SynthRead goes beyond a single number: you get sentence-level highlighting for long or complex lines, estimates of reading time, and flags for passive voice or jargon where enabled. Writers use it to hit an eighth-grade target for broad audiences, tighten grant or policy text, or verify that lesson materials suit students. In practice, SynthRead helps you align copy with audience literacy data rather than guessing, then you edit with those priorities in mind.

Yes. The Free plan lets you use core tools—including AI detection and SynthRead-style readability analysis—within limits such as 1,500 characters per request and three analyses per hour, which is enough to try workflows before upgrading. Paid tiers (Starter, Pro, Expert, and Enterprise) raise or remove those caps, add DeepScan for harder detection cases, unlock API access on qualifying plans, and include priority support where listed. Many teams start on Free to validate quality on real drafts, then move up when volume, batch jobs, or integrations require it. Pricing is published on the Pricing page with per-plan feature matrices so you can compare character budgets, rate limits, and API availability. Annual billing often reduces effective monthly cost if you commit for a year. If you are unsure which tier fits, estimate monthly words processed across tools and compare to the plan tables. In summary, you can ship real work on Free within the stated limits, then scale when your usage or compliance needs grow.

Open the Grammar Checker from the app menu, paste or type your text, and run a full check or a focused pass such as spelling-only or grammar-only. The tool surfaces correctness issues—misspellings, agreement, punctuation—and many clarity and style suggestions, such as wordiness or inconsistent tone, each tied to the exact span of text with a short explanation. You may accept or dismiss suggestions individually or apply grouped fixes when that fits your workflow. Multilingual support and English variants (for example U.S. versus U.K.) help global teams keep a consistent standard. Research on writing assistance consistently finds that combining automated feedback with human review yields better outcomes than either alone, which is why SynthQuery pairs grammar checking with readability metrics and, when needed, detection or plagiarism tools in adjacent tabs. For emails, essays, or support macros, run grammar first, then SynthRead if you need a target grade level. That sequence keeps correctness and audience fit explicit before you publish.

Most SynthQuery tools accept plain text pasted into the editor, which is the most reliable path because hidden formatting from word processors cannot interfere with counts or detection. Several flows also support uploads such as PDF or DOCX for readability and detection where the product UI lists them, and URL ingestion where the page content can be fetched cleanly. Image-based tools—including the AI Image Detector and Background Remover—accept common raster formats (for example JPEG or PNG) with size limits shown beside the upload control. Video or audio-heavy workflows may require transcripts or specialized detectors when available. Because parsers differ by format, always check the on-page hint for maximum file size and supported extensions before batching work. If a PDF is scanned images only, you may need OCR elsewhere first. In short, match the input type to the tool, respect the stated limits, and prefer paste when you need pixel-perfect character counts or consistent detection baselines.

Yes. Pro-tier and higher plans include REST API access for teams that want to call detection, readability, humanization, plagiarism, summarization, translation, and related endpoints from their own applications or CI pipelines. Authentication, rate limits, and payload schemas are documented on the API Documentation page alongside example requests. This pattern suits product companies automating moderation, agencies batch-checking client copy, and enterprises integrating checks next to a CMS or learning system. Machine-readable access lets you log usage consistently and enforce internal policies without manual copy-paste. When designing integrations, budget for the same character limits as the web app unless your contract specifies otherwise, and cache idempotent results where policy allows to stay within quotas. For security, rotate API keys, use HTTPS, and avoid sending highly sensitive fields you are not required to process. Overall, the API mirrors core product capabilities so you can embed SynthQuery where your authors already work instead of treating analysis as a separate step.