Directory decisions
Recently Not Listed
Not listed means AgentDish passed on publishing the submission in the directory. It is not a claim about the company or product. See the Terms of Service.
Show HN: Freightfloor. Have 3 minutes? Drive a forklift vs. AIThe page is too sparse to support a useful public listing. The crawl only exposes the title and a short brand line, with no description of the product, features, use case, examples, or AI functionality.
Show HN: Cirkit – Annotate, Trace, and Ask AI About Schematic PDFsThe crawl snapshot is too thin to support a public listing. The page only exposes the name and a marketing description in metadata, but the visible page text does not show product functionality, screenshots, workflow, pricing, docs, or any other concrete evidence that visitors can evaluate.
Show HN: Harness Corp – An incremental game powered by a local LLM using WebGPUThe snapshot is too thin to support a useful public listing. It only shows the page title and no description, features, screenshots, instructions, or evidence of how the LLM or WebGPU setup works.
Show HN: Agent-B.V., adversarial tests and on-chain reports for AI agentsThe page is too thin to support a useful public listing. The crawl snapshot only shows a title and one line of text, with no workflow, screenshots, examples, pricing, docs, or explanation of how the AI-agent verification and escrow tools actually work.
Show HN: Nekot, an iPhone app that warns you before Claude Code limits hitOnly Claude Code and Codex are explicitly supported in the snapshot. No deeper technical docs, API details, or screenshots of the full workflow are v Android is mentioned as coming soon, but there is no live version shown. The page appears to be placeholder or unfinished content.
Show HN: An animated map of the people and money in AIThe visible snapshot does not show concrete entries or named examples No details on data coverage, update cadence, or verification process Missing information on available interactions such as search, filters, or export The crawl found too little product detail for a reliable listing.
Show HN: Arketum – MCP server that makes coding agents stop for human approvalThe snapshot is effectively inaccessible: the page returns 403 and only shows “Just a moment...”, with no description, usage, features, or proof that this package is an AI product beyond the submitted title.
Show HN: Sectr – Serverless backend platform for easy deployment of AI agentsNo pricing or plan details beyond a beta/free mention No documentation pages or API reference visible in the snapshot No evidence of production scale, uptime, or supported cloud regions The page appears to be placeholder or unfinished content.
Show HN: Make your AI agent think before it codesNo product landing page, pricing, or team information is visible The crawl does not show screenshots or demo output Maintenance status beyond commits and stars is hard to judge from the snapshot The page appears to be placeholder or unfinished content.
Show HN: Agent-relay for multi-vendor coding sessionsNo clear installation or dependency section beyond example commands No visible releases, usage stats, or maturity signals in the snapshot The full protocol and adapter details are only partially visible from the crawl The page appears to be placeholder or unfinished content.
Discovering and exploiting vulnerabilities in real-world software with ClaudeThis is a research article, not a product/tool/page with enough listing value for AgentDish, and the snapshot does not present a usable AI product, API, repository, or agent offering. It is informative and clearly AI-related, but it falls below the threshold for a public tool listing.
The Agent SDKNo pricing or plan details shown Limited code example and no full API reference in the snapshot Supported integrations and constraints are only partially described The page appears to be placeholder or unfinished content.
AIReiterOnboarding and auth details are not fully visible No error-handling, limits, or SDK details in the crawl Some page content is truncated near the end The page appears to be placeholder or unfinished content.
The Exponential Time Hypothesis Is False [pdf]The page is just a file view for a PDF on GitHub, and the snapshot does not reveal the document’s contents, purpose beyond the title, or any AI tool, agent, API, or usable project details.
Show HN: Watch GPT-2 compute your sentence, one matrix at a time, in 3DThe page appears to be a thin demo landing page with almost no crawl-visible content beyond the title and meta description. It suggests an interesting GPT-2 visualization, but there is not enough on-page evidence about how it works, what users can do, or whether it is a substantial public product listing.
Show HN: Give our AI coding runner a real task and watch it until mergeThe broader platform is only lightly described; visitors cannot tell the full fe Pricing and plans are mentioned in navigation, but the actual offer is not visib The live page is centered on one project, so it is not obvious how reusable the The page appears to be placeholder or unfinished content.
Show HN: The Large Concepts Model – A DAG-based neurosymbolic alternative to RAGThis is a research preprint, not a product page, and the snapshot does not show an actual tool, API, repository, or usable software artifact. The page explains an architecture idea in detail, but there are no implementation links, demo, installation steps, code, or evidence that visitors can try or adopt it as a product.
Show HN: EveryGen AI – AI images, short videos and copy rewrites in one tabPricing is only hinted at; plan details are not visible. The snapshot does not explain account limits, credits, or export restrictions. Differences between the listed models/tools are not fully described on-page. The page appears to be placeholder or unfinished content.
Harness-aware evaluation of plugins (OpenAI Cookbook)This page is useful developer documentation, but the crawl snapshot does not show enough standalone product substance for a directory listing. It reads as a docs entry inside a large navigation-heavy site, and the visible content does not surface the actual cookbook example, workflow, code sample, or concrete steps beyond the title and surrounding index.
Show HN: Skill-spotter, auto detect agent prompts and commands to create skillsNo visible evidence of adoption beyond a small star count. No live demo or independent screenshots visible in the crawl snapshot. Performance and reliability claims are not backed by benchmarks or release histo The page appears to be placeholder or unfinished content.
Cognitive Mercantilism: AI panic resembles fears of early capitalismThis is a commentary essay, not an AI tool, product, API, repository, or other directory-worthy AI project. The page discusses AI politics and economic analogy, but offers no software, service, or builder-facing utility to list.
Pusaka Intelligence – AI equity research on US large capsOnly one sample report is visible, so the depth and consistency of coverage acro The methodology is described at a high level, but the valuation model details ar Several plan features are marked 'coming soon,' making current tier differences The page appears to be placeholder or unfinished content.
AI coding agents hijacked using AGENTS.md file in Downloads folderThis is a news article, not an AI tool, product, API, repository, or other listing-worthy AI project. The page documents a security issue involving AI agents but does not present a buildable or usable product/page for the directory.
Show HN: Amni-CAD – free browser CAD with FEA that Claude can drive over MCPNo screenshots, video, or interactive demo are visible in the crawl snapshot. The page lists many capabilities, but the snapshot does not show real examples o The free-vs-paid boundary is described only at a high level, so the practical li The page appears to be placeholder or unfinished content.
Show HN: God – real-time control room for LLM agent societies (EMNLP 2026 demo)The visible snapshot is truncated, so some implementation and deployment details The page does not show enough to confirm all supported backends or integrations The page appears to be placeholder or unfinished content.
How we test LLM serving changes for coding agents, two hours of GPUs per changeThis is a blog post, not a product or tool listing. It does describe Modelplane and LLM serving experiments, but the page is primarily an editorial writeup and does not provide enough standalone product evidence for a useful directory entry.
Show HN: A 3D mesh playground for coding agents (three.js + manifold-3d)The page is too thin to support a public listing: the crawl snapshot only shows a title and a fragment of text, with no description, features, examples, docs, or clear product usage.
Show HN: Carve a PumpkinThin page with almost no evidence of a public AI product, agent, API, or developer tool. The snapshot looks like a personal blog post for a seasonal game, not an AI-powered offering, and it does not provide enough information for a useful directory listing.
Show HN: Call4me – a ChatGPT plugin that lets it make phone calls on your behalfThe page is inaccessible in the crawl snapshot and provides no visible product details, usage, or evidence beyond the supplied title.
Show HN: Open-Source Cybersecurity Platform Uses AI Agents to Protect BusinessesThe crawl snapshot only shows a Cloudflare block on Medium, so there is no visible evidence of an AI product or enough page content for a public listing.
From Software Engineering to AI Research or a Tale of 413 "No"This is a personal career narrative, not an AI tool, product, API, repository, or developer resource suitable for an AI directory listing.
Reviewing the evidence on worker retraining programs Download the reportThis page is a research/report article from Anthropic, not an AI-powered product, developer tool, repository, API, or agent. It provides policy analysis and evidence review content rather than a tool or product visitors can use or evaluate as a listing.
Yeti, an MCP server that guides product design from requirements to JiraNo screenshots or visual proof of the dashboard UI No examples of real output from the MCP tools Limited evidence of usage, releases, or maintenance beyond the repository page The page appears to be placeholder or unfinished content.
Show HN: Aurora Agentic TerminalThe page is too thin to support a useful public listing. The crawl snapshot only shows a title and one line of visible text, with no feature details, screenshots, docs, pricing, setup instructions, examples, or explanation of how the terminal works in practice.
Show HN: Koala – Run AI Agents in Isolated Linux VMs on macOSPreview status only; stability and maturity are still limited No pricing or licensing details visible on the crawled page Support appears limited to Apple silicon macOS 26+ Full documentation and examples are referenced but not fully visible in the snap The page appears to be placeholder or unfinished content.
Show HN: Python script to evaluate your RAG pipeline against a golden setNo visible benchmark output, release notes, or test coverage. Limited evidence about support for larger workflows or export formats. Maintenance cadence is hard to judge from the snapshot alone. The page appears to be placeholder or unfinished content.
The attack chain behind the v1.107.0 LiteLLM releaseThis is an incident postmortem, not an AI tool, agent, API, repository, or product listing. While it is detailed and clearly about LiteLLM’s security breach, it does not present a usable AI product page for directory inclusion.
Show HN: Autonomous root for Tcl Android TV via CVE-2022-22706The snapshot is too thin to support a useful public listing. It only shows the repository name, a short exploit-chain description, and basic repo metadata, but not enough README, setup, proof-of-concept details, or usage context to judge the project meaningfully.
Show HN: Mark – AI chat that answers with diagrams, quizzes and 3D modelsThe page is too thin for a useful public listing: it gives a general promise of an AI chat assistant, but the crawl snapshot does not show enough product detail, features, workflows, pricing, docs, or examples to judge what users can actually do with it.
How to get picked for authentication by coding agentsNot a standalone software product or API listing The snapshot doesn’t expose the underlying replay board or full methodology deta Some of the page is framed as market analysis rather than a buildable AI tool The page appears to be placeholder or unfinished content.
New discoveries, old knowledge: AI finds connections humans missedThis is an editorial blog post, not a product, tool, API, repository, or service listing. The crawl snapshot does not show an AI offering a visitor could use or evaluate as a public directory entry.
Show HN: CATT 0.2.1: Test agents/services against a Muse/Instinct-alike, in CINo quickstart or installation instructions are visible in the crawl snapshot. The page does not show example command output or screenshots of a full run. Maintenance signals are limited in the snapshot beyond commit count and stars. The page appears to be placeholder or unfinished content.
Show HN: cotex, an Elixir-based Cursor on Target router AKA a TAK serverThe page is too thin to support a useful public listing. The crawl only shows the product name and a brief description, with no feature list, screenshots, docs, setup steps, examples, pricing, or proof of how the product works.
Abliteration – Removing LLM refusal behaviors via representation engineeringThe crawl snapshot shows a broken GitHub page with only a 404 message and no visible product, README, installation, examples, or usage details. There is not enough evidence for a useful public listing.
Show HN: Uberblok – an interactive museum of ZFS internalsThis is a polished educational project, but it is not an AI-powered product, AI agent, developer tool, API, or AI-adjacent project. The page is about ZFS documentation and interactive explanation rather than an AI tool, so it falls outside the directory’s target scope.
Using LLM for confidence scoring is a bad ideaThis is an informational blog article, not an AI product, tool, API, repository, or agent page with a public listing-worthy offering. The page explains an idea about LLM confidence scoring but does not present a product, implementation, demo, or reusable AI utility.
Show HN: Reverse Engineered Claude Code Desktop – Harness StudioNo installation or quick-start steps visible in the crawl snapshot No pricing or deployment model shown No demo media or screenshots visible in the crawl snapshot The page appears to be placeholder or unfinished content.
Show HN: Linden – a rule based verification layer for AI agentsThe crawl snapshot is effectively a placeholder page with only the text “frontend.” It does not provide enough visible evidence to support a public listing for an AI product or developer tool.