{"products":[{"name":"Vorlagenplatz","tagline":"Vorlagenplatz bietet kostenlose Vorlagen zum direkten Ausfül","description":"Kostenlose Vorlagen zum Ausfüllen & Ausdrucken\r\nFülle Vorlagen direkt online aus oder lade die leere Version als PDF herunter – ohne Anmeldung und komplett gratis. Von der Kündigung bis zur Bastelvorlage.\r\n\r\nVorlage suchen","website":"https://vorlagenplatz.de/","page":"https://ailaunch.dev/p/vorlagenplatz","slug":"vorlagenplatz","categories":["research","design"],"pricing":"Free","launchDate":"2026-09-30","upvotes":1,"comments":0,"featured":true,"hackerNewsPoints":null},{"name":"Mini-AGI","tagline":"Dynamic continual learning model trained on 8GB VRAM","description":"mini-AGI is a byte-level language model that trains from scratch on a single 8 GB VRAM GPU and keeps learning from a continuous stream of text. Weights live on disk and are paged onto the GPU as needed, so model size is bounded by disk spac","website":"https://github.com/volotat/mini-AGI/","page":"https://ailaunch.dev/p/mini-agi","slug":"mini-agi","categories":["infra","research"],"pricing":"Open Source","launchDate":"2026-09-23","upvotes":0,"comments":0,"featured":false,"hackerNewsPoints":277},{"name":"Cactus Needle 3","tagline":"8-29MB automation models can match DeepSeek V4 Flash","description":"Needle 3 is an 8-29 MB foundation model from Cactus for phones, wearables, robots, smart home devices, cars and microcontrollers. It trades general chat ability for strong tool calling, structured extraction and text embeddings, running ful","website":"https://cactuscompute.com/needle","page":"https://ailaunch.dev/p/cactus-needle-3","slug":"cactus-needle-3","categories":["infra"],"pricing":"Freemium","launchDate":"2026-09-24","upvotes":0,"comments":0,"featured":false,"hackerNewsPoints":236},{"name":"LLM Attention Visualization","tagline":"A visualization of the attention mechanism in LLMs.","description":"An interactive visualization of the attention mechanism in a small large language model. Hover or tap any generated token to see which earlier tokens influenced it, with opacity scaled by attention weight across all heads and layers. Exampl","website":"https://ishamf.dev/p/llm-attention-visualizer/","page":"https://ailaunch.dev/p/llm-attention-visualization","slug":"llm-attention-visualization","categories":["research","infra"],"pricing":"Free","launchDate":"2026-09-25","upvotes":0,"comments":0,"featured":false,"hackerNewsPoints":177},{"name":"JevBench","tagline":"A reproducible benchmark for typed decision models","description":"JevBench, from Benchmark Heaven, compares Jev-class decision models across intelligence, calibration, speed and cost. It sits alongside Benchmark Heaven's broader model rankings, which can be filtered by provider, lab, hosting region and da","website":"https://benchmarkheaven.com/jev-models","page":"https://ailaunch.dev/p/jevbench","slug":"jevbench","categories":["research","infra"],"pricing":"Free","launchDate":"2026-09-26","upvotes":0,"comments":0,"featured":false,"hackerNewsPoints":153},{"name":"Kinesis","tagline":"Control your Mac with the Meta Neural Band","description":"Kinesis is a highly experimental macOS app that lets you control your Mac with a Meta Neural Band. Swipe between desktops, open Mission Control, control music, and pinch and turn for volume or brightness. You choose the gesture mappings, pr","website":"https://github.com/callbacked/kinesis","page":"https://ailaunch.dev/p/kinesis","slug":"kinesis","categories":["productivity"],"pricing":"Open Source","launchDate":"2026-09-27","upvotes":0,"comments":0,"featured":false,"hackerNewsPoints":122},{"name":"TinyAIArena","tagline":"Watch AI agents battle it out","description":"Tiny AI Arena is a browser game where AI agents play against each other while you watch. You can step through a match frame by frame, auto-play it, start a new game, and follow the game log and global chat.\n\nFirst shared by its maker on Sho","website":"https://tinyaiarena.com/","page":"https://ailaunch.dev/p/tinyaiarena","slug":"tinyaiarena","categories":["agents"],"pricing":"Free","launchDate":"2026-09-28","upvotes":0,"comments":0,"featured":false,"hackerNewsPoints":120},{"name":"TinyBrains","tagline":"Build the smallest neural network that plays strategy games well","description":"TinyBrains is a ranked ladder for small neural networks that play strategy games. Train a model, write an adapter, upload two files, and get measured into a weight class starting at 8 KiB.\n\nFirst shared by its maker on Show HN: https://news","website":"https://tinybrains.dev","page":"https://ailaunch.dev/p/tinybrains","slug":"tinybrains","categories":["research"],"pricing":"Free","launchDate":"2026-09-29","upvotes":0,"comments":0,"featured":false,"hackerNewsPoints":116},{"name":"CUA-S1","tagline":"A System One Model for Computer Use","description":"Cua gives AI agents computers they can use. It provides open-source desktop automation drivers, isolated cloud desktops (Cua Fleets), local macOS VMs (Lume), CUA-S1 specialist models for computer-use decisions, and Cua Bench for evaluating ","website":"https://github.com/trycua/cua","page":"https://ailaunch.dev/p/cua-s1","slug":"cua-s1","categories":["agents","infra"],"pricing":"Freemium","launchDate":"2026-09-30","upvotes":0,"comments":0,"featured":false,"hackerNewsPoints":95},{"name":"Engrim","tagline":"A universal – local-first SQLite memory engine for AI CLIs","description":"Engrim is a local-first, project-scoped memory engine for AI coding agents, stored in a single SQLite file. It lets developers switch between tools like Claude Code, Cursor, Codex CLI, GitHub Copilot CLI, OpenCode and Google Antigravity on ","website":"https://github.com/timgordontg/engrim","page":"https://ailaunch.dev/p/engrim","slug":"engrim","categories":["coding","agents"],"pricing":"Open Source","launchDate":"2026-10-01","upvotes":0,"comments":0,"featured":false,"hackerNewsPoints":93},{"name":"FrontierHarness Eval","tagline":"Same model, 12 harness configurations, up to 17.5x cost difference","description":"FrontierHarness Eval compares coding-agent harnesses on identical software engineering tasks using the same model and runtime. Across 12 configurations of 9 harnesses, including Codex, Claude Code and OpenCode, pass rates were similar but c","website":"https://frontierharness.org","page":"https://ailaunch.dev/p/frontierharness-eval","slug":"frontierharness-eval","categories":["coding","research"],"pricing":"Open Source","launchDate":"2026-10-02","upvotes":0,"comments":0,"featured":false,"hackerNewsPoints":82},{"name":"PacBench","tagline":"How well can models one-shot a Pac-Man game?","description":"PacBench tests how well a model and harness can recreate Pac-Man from one prompt: \"Create a Pac-Man game in a single html page.\" Entries from different models and harnesses can be sorted by score, cost, time and tokens, with run data for ea","website":"https://jonclegg.github.io/pacman-bakeoff/","page":"https://ailaunch.dev/p/pacbench","slug":"pacbench","categories":["coding","research"],"pricing":"Free","launchDate":"2026-09-23","upvotes":0,"comments":0,"featured":false,"hackerNewsPoints":78},{"name":"Chess Post-mortem Skills","tagline":"Claude Code skills that turn your chess games into readable post-mortems","description":"Claude Code skills that turn one of your chess games into a post-mortem you can read: plain-language explanations of your mistakes, checked against Stockfish, an annotated PGN, an HTML viewer and a narrated video. You can also record yourse","website":"https://github.com/brumar/chess-postmortem-skills","page":"https://ailaunch.dev/p/chess-post-mortem-skills","slug":"chess-post-mortem-skills","categories":["coding","productivity"],"pricing":"Open Source","launchDate":"2026-09-24","upvotes":0,"comments":0,"featured":false,"hackerNewsPoints":75},{"name":"Lossless-memory","tagline":"A personal AI memory that never summarizes","description":"lossless-memory is a long-term memory system for a personal AI assistant that never summarizes. It keeps every raw line, timestamps everything, and searches by time first and words second, returning results in chronological order. A small i","website":"https://github.com/aru-labs/lossless-memory","page":"https://ailaunch.dev/p/lossless-memory","slug":"lossless-memory","categories":["infra","productivity"],"pricing":"Open Source","launchDate":"2026-09-25","upvotes":0,"comments":0,"featured":false,"hackerNewsPoints":67},{"name":"Jevstiller","tagline":"Distill Jev into a local model – with a disagreement bound","description":"Jevstiller distills a hosted Jev classifier into a local model, with a measured bound on how often the local model disagrees with Jev. It runs as a drop-in proxy, so batch jobs avoid a network call per answer while keeping answers close to ","website":"https://jevstiller.pages.dev/posts/the-guarantee/","page":"https://ailaunch.dev/p/jevstiller","slug":"jevstiller","categories":["infra"],"pricing":"Open Source","launchDate":"2026-09-26","upvotes":0,"comments":0,"featured":false,"hackerNewsPoints":63},{"name":"Pizza Bot","tagline":"An inbox for AI agents that work in the background","description":"Pizza Bot is an inbox for long-running AI work. Start or schedule a task, walk away, and come back to finished work in Unread and decisions waiting in Action. Agents keep running when you disconnect, and the same app works on desktop, web a","website":"https://github.com/pizza-bot-app/pizza-bot","page":"https://ailaunch.dev/p/pizza-bot","slug":"pizza-bot","categories":["agents","productivity"],"pricing":"Open Source","launchDate":"2026-09-27","upvotes":0,"comments":0,"featured":false,"hackerNewsPoints":61},{"name":"MultiMatte","tagline":"A Promptable Image Background Removal Model","description":"MultiMatte is a background removal model you steer with words: name the object you want to keep and it removes everything else. It is built on Meta's SAM 3 with low-rank fine-tuning of 2.27% of its weights and produces alpha mattes, so fine","website":"https://usefeyn.com/blog/multimatte/","page":"https://ailaunch.dev/p/multimatte","slug":"multimatte","categories":["image","design"],"pricing":"Free","launchDate":"2026-09-28","upvotes":0,"comments":0,"featured":false,"hackerNewsPoints":57},{"name":"Doom or Bloom","tagline":"Map your AI worldview","description":"Doom or Bloom lets you map your worldview on how AI will change the future and compare it with public figures. Answer a few questions to see where you land between doom and bloom, and between incremental and civilizational change.\n\nFirst sh","website":"https://www.doom-or-bloom.com","page":"https://ailaunch.dev/p/doom-or-bloom","slug":"doom-or-bloom","categories":["research"],"pricing":"Free","launchDate":"2026-09-29","upvotes":0,"comments":0,"featured":false,"hackerNewsPoints":56},{"name":"Ordewell","tagline":"Turn one goal into an ordered plan of coding-agent tasks","description":"Ordewell turns one goal into an editable plan of coding-agent tasks. A read-only planner researches your repository, asks about anything unclear, and returns a dependency graph where each task names its agent, model and thinking effort. Ind","website":"https://github.com/ordewell/ordewell","page":"https://ailaunch.dev/p/ordewell","slug":"ordewell","categories":["coding","agents"],"pricing":"Open Source","launchDate":"2026-09-30","upvotes":0,"comments":0,"featured":false,"hackerNewsPoints":56},{"name":"Masala (formerly Panel)","tagline":"A research workspace where the agent can build its own panes","description":"Masala, called Panel until September 2026, is a research workspace where an AI agent works beside you. Chat, files, PDFs and notebooks sit in one dock, and the agent can create custom viewers and apps when it needs them. It's an early build","website":"https://github.com/greentfrapp/panel","page":"https://ailaunch.dev/p/masala","slug":"masala","categories":["research","agents"],"pricing":"Open Source","launchDate":"2026-10-01","upvotes":0,"comments":0,"featured":false,"hackerNewsPoints":54}],"periods":["today","week","month","all"]}