{"slug": "show-hn-slopsift-a-local-graph-backed-linter-for-ai-writing", "title": "Show HN: Slopsift – a local, graph-backed linter for AI writing", "summary": "SlopSift, a local graph-backed linter for AI writing, uses a custom-trained dependency parser to detect canned arguments, unsupported claims, and filler in text. The tool runs on-device via quantized ONNX weights in Node and browser WebAssembly, avoiding API calls. It outputs findings in ESLint-shaped JSON for CI integration.", "body_md": "### Small enough to ship\n\nQuantized ONNX weights run locally in Node and browser WebAssembly.\n\nNatural-language processing. Local first.\n\nSlopSift uses a small dependency parser we trained to map the relationships between words and find canned arguments, unsupported claims, and filler.\n\nLoading the on-device NLP model…\n\nMore than a word list\n\nMany writing tools stop at word matching and basic parts of speech. SlopSift follows the relationships between words, so rules can recognize how a sentence makes its claim—not only which vocabulary it uses.\n\nOur custom-trained compact model maps tokens, parts of speech, and the grammatical relationships holding the sentence together.\n\nAuthorable rules inspect the graph for structural tells. Every finding names what matched and the exact text that triggered it.\n\nErrors are strong tells. Warnings need attention. Notes are candidates—not a machine pretending to know who wrote the sentence.\n\nNot an API wrapper\n\nSlopSift starts with a compact pretrained English encoder and trains it for parts of speech and dependency parsing. Training combines structured distillation from a larger parser with 50 controlled examples targeting grammatical relationships used by the linter. We reserved separate template families for evaluation.\n\nQuantized ONNX weights run locally in Node and browser WebAssembly.\n\nUPOS, dependency arcs, and dependency relations—not a generic text score.\n\nThe model and deterministic rules run on-device. No remote judge reads the text.\n\n“As an AI language model...”\n\nThree paragraphs use the same canned outline.\n\nAn actorless passive may be hiding responsibility.\n\nGlob files, lint Markdown, inspect code comments, and emit ESLint-shaped JSON in CI.\n\nLet coding agents run the real linter, interpret its findings, and edit without flattening your voice.\n\nYes, AI helped build this.\n\nThat is the point. SlopSift is not an AI detector and it does not pretend to know who typed a sentence. It catches vague or inflated writing. It also catches repetition and borrowed certainty. Human beings do those things too.\n\nUse `--format json`\n\nfor machines, `--level info`\n\nfor the full suspicious pile.", "url": "https://wpnews.pro/news/show-hn-slopsift-a-local-graph-backed-linter-for-ai-writing", "canonical_source": "https://slopsift.dev/", "published_at": "2026-07-18 14:03:00+00:00", "updated_at": "2026-07-18 14:21:12.821563+00:00", "lang": "en", "topics": ["natural-language-processing", "developer-tools", "ai-tools"], "entities": ["SlopSift", "ONNX", "WebAssembly"], "alternates": {"html": "https://wpnews.pro/news/show-hn-slopsift-a-local-graph-backed-linter-for-ai-writing", "markdown": "https://wpnews.pro/news/show-hn-slopsift-a-local-graph-backed-linter-for-ai-writing.md", "text": "https://wpnews.pro/news/show-hn-slopsift-a-local-graph-backed-linter-for-ai-writing.txt", "jsonld": "https://wpnews.pro/news/show-hn-slopsift-a-local-graph-backed-linter-for-ai-writing.jsonld"}}