Discover the Best AI Tools

Collects news and publishes source-linked, game-style planetary patch notes with an RSS feed. New editions require Groq credentials and an archive secret; ordinary visits read saved editions, and generated labels may be wrong.

Goutoujunshi Jev Chat previews chat capture, user-checked transcripts, intent analysis and reply drafts across Mac, Windows and Android. Sending remains manual. Android currently cannot capture WeChat, Windows and Android need further device validation, and Mac ships as source. Configured cloud OCR/models receive relevant content; displayed weights are not verified intent or relationship-success probabilities.

Jev Chat Jarvis for Apple Silicon macOS reads supported chat windows, estimates intent and ranks reply drafts; sending remains manual. Judgment can be local or cloud, but generation and configured relays can receive chat history. It requires permissions, has limited tested layouts, and local judgment needs sufficient memory. Scores are model estimates, not a person's true intent.

JevChat Windows reads supported WeChat and KakaoTalk windows with local OCR, uses Jev and a text model to rank reply drafts, and can fill the input box; sending remains manual. It needs Windows 10 1903+ or 11 and provider keys. Chat text reaches external models; OCR and intent scores can be wrong, and releases are unsigned.

A prose linter combining static rules with probabilistic checks for clarity and writing habits across documents and webpages. Semantic checks send text to a configured decision model; alternative models need threshold tuning, while static-only mode works without a model key.

Slop Grader evaluates text against line-level and document-level rules with Jev and produces reports for human or agent revision, with incremental caching. It requires TypeSafe or OpenRouter access; exhaustive rule coverage does not eliminate false positives or make it an authorship detector.

Sniff Test checks Markdown and plain text against configurable house-style rules and reports flags with locations and confidence. Regex rules run locally; judgment rules send paragraphs to TypeSafe after consent. It flags issues without rewriting the draft.

AI News Aggregator is a runnable pipeline that collects RSS, research and social sources, ranks and summarizes coverage with agents, and publishes briefings through a web frontend and feeds. It requires model-provider configuration; some social sources also need API credentials.

Writing Style Checker flags passive voice, vague wording, repetition and patterns common in AI-generated prose through an editor, API or CLI/MCP. Its explainable pattern checks suggest edits; they cannot prove whether a person or AI authored the text.

judging-with-typesafe includes Python tools for typed Jev queries and paragraph-level text review using criteria and uncertainty bands. It sends selected paragraphs to TypeSafe and writes scores to JSON, requiring an API key. It can help triage claims for human review, but cannot verify facts or sources; thresholds need evaluation for the material.

A local browser and CLI tool that generates Korean prompts for incident triage, deployment checks and coding-agent handoffs. The free generator is separate from its paid template pack; it does not itself investigate incidents or execute fixes.

patina uses deterministic rules to revise AI-sounding prose in Korean, English, Chinese and Japanese, with change explanations and fidelity checks. It offers CLI, skill and web interfaces. Web rewriting runs on the server; editing scores are not proof of authorship.

Agent skill that turns ideas or story outlines into text-to-video prompts after assessing their suitability. Requires a compatible language-model agent; it generates prompts rather than video, and video rendering needs a separate model or service.

Deterministic Python CLI/MCP flags configured filler phrases and produces repeatable scores and CI exit codes without a model or API key. Supports custom phrase lists; matches are editorial cues, not authorship detection or proof a phrase can always be deleted.

Local CLI/MCP and CI linter flags prose rhythm and phrasing patterns without model calls, with a voice layer for drafting assistants. It identifies style issues, not authorship; reliability varies by writing register and email calibration is weak.

Provides media-model catalog search, model guidance and prompt recipes through MCP, handing users off to Kavel via generated URLs. It does not execute image or video generation inside MCP. The connector needs no account or API key, while use of the target service has separate conditions.

Local CLI and MCP tools for detecting and removing supported chat-copy artifacts and comparing factual tokens before and after edits. Russian is the default profile, with English cleanup support. Regex findings do not establish AI authorship or semantic truth; markup-sensitive modes and manual review remain necessary.

An MCP proposal workflow tool for freelancers and consultants that manages a local library of past proposals and helps an AI client analyze briefs, draft proposals and prepare follow-up writing. Drafts use saved examples as references and remain dependent on the connected AI client; proposal content used for drafting is shared with that client’s model, not kept entirely offline.

Prose-editing CLI and agent plugin with offline regex cleanup and optional model-assisted rewriting; it offers style adjustments without guaranteeing detector outcomes or factual preservation.

A text rewriting pipeline that combines LLM rewrites and machine-translation stages, with documented intermediate steps and local execution.

InkOS is a Story Creation AI Agent designed for novels, scripts, translation, interactive games, and IP content.

OpenDraft is an open-source AI tool for writing research papers that includes citation verification against CrossRef, OpenAlex, and Semantic Scholar, with export options for PDF, Word, and LaTeX.

MCP server for full control of Apple Keynote through AppleScript automation, allowing users to create, edit, and export presentations via natural language.

Converts any URL to LLM-ready Markdown using real Chrome browsers.