83.7K
The open-source CapCut alternative
Tools Bench.
Product launches and open-source repos with enough signal to earn a second look.
Last Brew Time: Aug 16, 2026, 10:56 AM PT
Sources
GitHub
72.4K
Local UI to run and train LLMs and diffusion models, including Qwen3.8, Kimi K3, MiniMax-H3, Gemma 4, DeepSeek-V4, FLUX and more.
39.9K
ToolJet is the open-source foundation of ToolJet AI - the enterprise app generation platform for building internal tools, dashboard, business applications, workflows and AI agents 🚀
25.3K
Beautiful, Modern & Opinionated Linux
14MB foundation model for tiny devices; phones, wearables, smart home, and robots.
Meta-Framework of Spatiotemporal Composability
Product Hunt
Inferock-bench is a local proxy that sits between your app and OpenAI, Anthropic, Gemini, or OpenRouter shaped calls. It captures per-call token usage, failures, and retries, then generates an independent receipt showing what you were billed and how much you're actually overpaying for.
231
GLM-5.3 is Z.ai's latest model built for complex, long-horizon coding tasks. Through massive post-training scaling, it achieves open-source SOTA in agentic coding and demonstrates emergent capabilities in vulnerability discovery and cyber defense.
178
Presidents and CEOs get strategy teams. You get… doomscrolling? Meet Zetik: an AI agent team running a full intelligence cycle — collect, filter, analyze, brief — across podcasts, papers, code, tweets & news, 24/7. It tracks whatever matters to you, in near real time.
154
Attyn brings intelligence to the cursor inside the apps you already use. Rewrite selected text with Inline Assist, speak finished words with Realtime Dictation, ask about what is on screen with Screen Assist, and turn questions into visual explanations with Blackboard. Use Attyn Credits, bring a provider key, or run a supported local model. Attyn is bootstrapped and independent. Available for macOS today, with Windows on the way. Every new account starts with 500 launch credits.
133
nenspace. an extended mind, not a second brain: your own, made larger. the lo-fi of LLMs fused with the one space that catches your thoughts and cultivates them: working memory, tasks, notes, habits, logbook. AI-fused, but not forced. your head holds about four things. nenspace holds the rest. free to start · web + ios
AO helps you manage all your coding agents in a single place. Monitor your whole agent fleet on a kanban view as each task goes from working to PR to Running tests to Code Review. Plan the roadmap with your project specific orchestrator. Let the orchestrator breakdown tasks, delegate them to isolated agents, and manage the fleet for you. Stop tab switching between agent terminals. Start orchestrating. Free and open source under Apache 2.0.
YC Launch
A mobile app to watch vertical anime shorts, created with AI. MOCHI.TV · Summer 2026 · Consumer Tags: Artificial Intelligence, Consumer. Website: https://www.mochi.tv/
World Models for Accurate Humans Familiar · Summer 2026 · B2B Tags: B2B, Media, Advertising, AI. Website: https://thefamiliarlab.com
Trade Anything, Anywhere - 2,700 traders have traded $1.35B in 7 months Arbital · Summer 2026 · Fintech Tags: Artificial Intelligence, Fintech, Crypto / Web3, Investing, Trading. Website: https://www.arbital.xyz
AI agents that automate the complex manual workflows governments run on. Stratum Industries · Summer 2026 · Government Tags: Artificial Intelligence, GovTech, Automation. Website: https://stratumindustries.co/
Peer brokers every load autonomously, from quote to delivery. Peer · Summer 2026 · B2B Tags: Digital Freight Brokerage, Logistics. Website: https://peer-freight.com
Making the grid flexible to unlock GW power, enabling on-grid interconnection for data centers NODI Energy · Spring 2025 · Industrials Tags: Energy Storage, Hard Tech, Hardware, Energy, Electronics. Website: https://www.nodienergy.com
Hacker News
Hey HN, Henry from Cactus here! We previously released Cactus Needle, a 14MB agentic LLM for tool call, device use, and structured extraction for phones, wearables, smart homes, small robots and microcontrollers. We got really great feedback here, and have now incorporated the suggestions to release Needle 2. The whole model is a single 14MB binary that runs a full session in 28MB of RAM; 45m parameters at 2bit compression. Needle hits 500 tokens/sec decode speed on a Raspberry Pi 5, sits betwee... (530 points, 182 comments).
Doing research with agents is fun until they blow way past budget, jumble the sources, and don't even give you the best possible answer, just sound confident. And if you want to run some research task on local data - you have no idea where your data ends up after the prompt consumes it. So I built this tool: a deep-research agent with an enforced budget, verified quotes, and a privacy boundary for local data. 1. Never spend more than you budgeted (measured overshoot is 0%). 2. Every claim carrie... (98 points, 14 comments).
Hacker News discussion (7 points, 1 comments).
That's a bold claim. But I genuinely feel like I might have actually solved computer use (demo: https://x.com/mdlahfir/status/2088109763783700827?s=20 ) For context, I've been building agent-desktop (Inspired by agent-browser by Vercel Labs), an automation CLI for desktop apps. It's like Playwright but for desktops, not just native, but for Chromium apps as well. Trust me, yes, Chromium apps whose accessibility tree is dense. MacOS is GA; I'm almost close to launching for Windows and Linux! So,... (6 points, 0 comments).
We just released the first public results from the Agent Memory Leaderboard (AML). The first evaluation focuses on Text Memory across two tracks: - Open-source Methods - Commercial Products 136 teams registered, and 69 representative memory frameworks completed the first evaluation. Commercial Products — Text Memory: 1. MemoraX — 58.02 2. MemOS — 45.89 3. NTES-MEMORY-SMART — 44.21 The benchmark uses a common evaluation framework and a clearer system boundary: Memory system: Add → Search Benchmar... (3 points, 1 comments).
Hi HN, I built the first version of Mocktail years ago. I recently came back to the project, and after a pretty substantial rebuild, v4 is now out. Mocktail is a free and open-source, self-hosted mock API server with a built-in dashboard and database, packaged as a single ~25 MB binary. You can run it locally or on your own infrastructure — no account or hosted service required. You can define endpoints and responses, generate realistic data per request, customize headers, status codes and laten... (22 points, 1 comments).
HF Spaces
2.6K likes
Demo of the Collection of Qwen Image Edit LoRAs Qwen-Image-Edit-2511-LoRAs-Fast is a Hugging Face Space tagged with gradio, mcp-server, region:us. It has 2576 likes on Hugging Face.
441 likes
Unified memory evaluation · Results expected August 12. Agent Memory Leaderboard · 记忆之巅 A unified, open, and reproducible evaluation platform for long-term memory systems and memory-enabled agents. Agent Memory Leaderboard (AML) compares research methods and commercial products under one evaluation contract. Candidate systems implement memory Add and Search; the official platform fixes Answer, Eval, datasets, models, configurations, result review, and publication. > First public release: The inaugural verified leaderboard is expected to be published on August 12, 2026. > 首期发布： 首期经核验榜单预计将于 2026 年 8 月 12 日发布。 Results are separated along two independent dimensions. Textual and coding tasks use...
257 likes
Video generation with a synchronized soundtrack MiniMax-H3 — unquantized, split across two Spaces Joint video and soundtrack out of a single denoising pass, at bfloat16 with no quantization anywhere. This Space is the denoising half: the 61.73 GiB transformer and the two autoencoders. The 62.14 GiB Qwen3-VL conditioner runs in qwen3vl-conditioner, which this Space calls over the gradio API for every request. The weights are the public MiniMaxAI/MiniMax-H3 diffusers checkpoint. MiniMax-H3 is 195.9 GiB in bfloat16 and a ZeroGPU Space is evicted at 150 GB of storage. An unquantized single Space is therefore impossible, which is why quantized demos of it run NVFP4 or float8 weights. Cut the Mini...
166 likes
Video generation with a synchronized soundtrack MiniMax-H3 — unquantized, split across two Spaces Joint video and soundtrack out of a single denoising pass, at bfloat16 with no quantization anywhere. This Space is the denoising half: the 61.73 GiB transformer and the two autoencoders. The 62.14 GiB Qwen3-VL conditioner runs in qwen3vl-conditioner, which this Space calls over the gradio API for every request. The weights are the public MiniMaxAI/MiniMax-H3 diffusers checkpoint. MiniMax-H3 is 195.9 GiB in bfloat16 and a ZeroGPU Space is evicted at 150 GB of storage. An unquantized single Space is therefore impossible, which is why quantized demos of it run NVFP4 or float8 weights. Cut the Mini...
127 likes
MiniMax Music 3 Studio — diffusers demo Streams full songs from lyrics + a structured caption using the MiniMaxMusic3Pipeline diffusers port. The input surface is a single Suno-inspired custom gr.HTML composer (Simple ↔ Studio modes, section-tag chips, structured-caption fields per the official prompting guide) that drives Gradio events via trigger()/props.value; styling uses only theme CSS vars so it follows the Citrus theme natively. Weights: MiniMaxAI/MiniMax-Music3 AoTI kernels: diffusers-internal-dev/MiniMax-Music3-aoti (compiled on RTX Pro 6000, matching ZeroGPU hardware) Generation streams chunk by chunk with a configurable playback headroom. The 8B language-model stage runs eager on....
108 likes
Ultra-fast local NVFP4 video + synchronized audio generation MiniMax-H3 Ultra Fast — local conditioner + pruned NVFP4 on Blackwell Joint video and synchronized sound from MiniMax-H3, rebuilt for a single 96 GB Blackwell ZeroGPU worker. | layer | optimization | |---|---| | Weights | 12.5 GB pruned NVFP4 transformer: 20.1B effective parameters instead of 33.1B/61.7 GiB BF16. | | Compute | Native CUDA 13 NVFP4 tensor-core GEMMs through comfy-kitchen; higher-precision norms, embeddings and output heads. | | Residency | Transformer, conditioner and both VAEs remain GPU-resident during generation—no layerwise CPU offload. | | Conditioner | Local 15.7 GB Qwen3-VL NVFP4-AWQ checkpoint containing onl...