Data/All products
Search and sort every AI product collected from the connected sources. Each row keeps its numbers in the source's own unit. Open a row to see its tags, timeline and original listings.
Products in view
391
Matching your filters, last 30 days
Classified
3,267
of 3,267 collected products
Awaiting classification
0
Picked up by the next pipeline run
Sources collecting
5/6
Product Hunt not connected
Filters apply to the table below and stay in the link when you share it.
Ranked by attention: votes, points or stars compared only within the same source, then averaged. Products with no source metric are left out.
| # | Product | Technology | Industry | Traction signal | Launched | |
|---|---|---|---|---|---|---|
| 1 | A AgentRun AgentRun is an open-source self-hostable managed agents platform with self-improving agents, sandboxes, tools and integrations, and behavior analytics. | AI Agents | Developer Tools | 99 / 100 138 points on Hacker News | Nov 12, 2026 | |
| 2 | O Open-source tamserve is open-source real-time object tracking for live video feeds that uses Meta's SAM 3 with a lightweight tracker to find and follow named objects. | AI Infrastructure | Logistics | 99 / 100 120 points on Hacker News | Nov 16, 2026 | |
| 3 | A A language model that lives in a boot sector A tiny char-level neural network with a 510-byte boot-sector inference engine that generates Linux boot messages. | AI Infrastructure | Unknown | 99 / 100 109 points on Hacker News | Oct 31, 2026 | |
| 4 | S Stroq Stroq provides a local audit showing what a coding agent did and why. | AI Infrastructure | Developer Tools | 99 / 100 96 points on Hacker News | Nov 5, 2026 | |
| 5 | W We compressed a 31B video+audio diffusion model to 14B The product compresses a 31B video and audio diffusion model to a 14B 6-step INT8 version. | AI Infrastructure | Unknown | 98 / 100 84 points on Hacker News | Nov 7, 2026 | |
| 6 | K Kapso An open-source engine from Leeroo that runs experiments, deploys the best solution, and learns from enterprise data work. | AI Agents | Enterprise IT | 98 / 100 61 points on Hacker News | Nov 12, 2026 | |
| 7 | T TrnRT TrnRT is an open-source LLM inference engine for AWS Trainium that serves models like GLM 5.3 flash with an OpenAI-compatible API. | AI Infrastructure | Developer Tools | 98 / 100 54 points on Hacker News | Oct 30, 2026 | |
| 8 | G Gauge Gauge is an open-source CLI that recursively improves docs, Skills, and other agent-facing surfaces by running coding-agent tasks and hillclimbing variants. | AI Agents | Developer Tools | 98 / 100 43 points on Hacker News | Nov 4, 2026 | |
| 9 | L llama-modchip llama-modchip runs large DeepSeek and MoE models on 64GB CPU with no GPUs and reports inference throughput. | AI Infrastructure | Developer Tools | 97 / 100 1,386 stars on GitHub | Oct 29, 2026 | |
| 10 | W Watch weight-space programming in action with WebGPU Interactive WebGPU walkthrough showing editing 0.016% of Qwen3.5-0.8B weights to reduce reasoning loops without gradient descent. | AI Agents | Developer Tools | 97 / 100 38 points on Hacker News | Nov 19, 2026 | |
| 11 | I inferssd Lets users run MoE models larger than RAM or VRAM by streaming experts from SSD, built on llama.cpp. | AI Infrastructure | Developer Tools | 97 / 100 1,262 stars on GitHub | Oct 22, 2026 | |
| 12 | G GPUWorks GPUWorks is a 40 MB GPU cluster manager for running vLLM/SGLang without Kubernetes. | AI Infrastructure | Developer Tools | 97 / 100 30 points on Hacker News | Nov 4, 2026 | |
| 13 | A Automatic deduction of FlashAttention kernel from ~25 lines of SQL The product automatically deduces a FlashAttention kernel from vanilla attention SQL using array, differentiation, and planner extensions. | AI Infrastructure | Developer Tools | 97 / 100 28 points on Hacker News | Oct 30, 2026 | |
| 14 | L lot-map lot-map is a local-only live map beside Claude Code chat that shows what the agent is doing including tool calls, model steps, hooks, skills and subagents. | AI Infrastructure | Developer Tools | 96 / 100 1,030 stars on GitHub | Nov 9, 2026 | |
| 15 | M Maple Maple is a pruning method to run Gemma 4 26B at 45 TPS on a 6GB GPU. | AI Infrastructure | Developer Tools | 96 / 100 21 points on Hacker News | Nov 8, 2026 | |
| 16 | F Flide Flide is a high-performance exact inference engine for Jev-like decision models installable via pip. | AI Infrastructure | Developer Tools | 96 / 100 126 points on Hacker News | Oct 24, 2026 | |
| 17 | L LLM weight editing tool made by a 15-year-old It is a tool for directly editing LLM weights, also called Laplace's Scissors. | AI Infrastructure | Developer Tools | 95 / 100 14 points on Hacker News | Nov 7, 2026 | |
| 18 | C Cryo Cryo is a programming language and toolchain for durable long-running workflows and AI agents that freeze and thaw. | AI Agents | Developer Tools | 95 / 100 12 points on Hacker News | Oct 27, 2026 | |
| 19 | M Markdown is 40× more token-efficient with SQLite Native provides open-source self-hosted context vaults using Markdown folders or SQLite projections for team messaging, docs and projects. | AI Infrastructure | Developer Tools | 94 / 100 11 points on Hacker News | Oct 27, 2026 | |
| 20 | A agent-hud-open Agent-hud-open is a HUD for Claude Code, Codex and 25 more coding agents that shows quotas, tokens and sessions and lets users answer requests in one click. | AI Infrastructure | Developer Tools | 94 / 100 650 stars on GitHub | Oct 22, 2026 | |
| 21 | M Maglev Maglev is described as an OpenRouter-like service for agent tools. | AI Agents | Developer Tools | 94 / 100 10 points on Hacker News | Nov 19, 2026 | |
| 22 | L Log every LLM call with tokens Llmtee is a lightweight LLM proxy that logs calls with tokens, cost, latency, and time to first token to SQLite or Postgres. | AI Infrastructure | Developer Tools | 94 / 100 10 points on Hacker News | Nov 11, 2026 | |
| 23 | T TheoremSmith TheoremSmith is a pipeline that mass-produces hard formal proof problems as RL training environments for LLMs. | AI Infrastructure | Developer Tools | 94 / 100 10 points on Hacker News | Nov 10, 2026 | |
| 24 | W WebsiteBench WebsiteBench is a benchmark testing whether AI agents can rebuild websites through browser exploration. | AI Agents | Developer Tools | 93 / 100 554 stars on GitHub | Nov 4, 2026 | |
| 25 | K KradleBench A multi-agent benchmark that tests AI collaboration, conflict and reaction in dynamic real-time Minecraft worlds. | AI Agents | Developer Tools | 93 / 100 9 points on Hacker News | Nov 18, 2026 |