Tools Bench.

Product launches and open-source repos with enough signal to earn a second look.

Last Brew Time: Oct 5, 2026, 11:15 AM PT

Insight

AI builders this run are converging on making the agent's surroundings into the product, turning agent infrastructure into the real battleground

01.PERSISTENT MEMORY

02.DETERMINISTIC GUARDRAILS AND VULNERABILITY HUNTING

03.ONE-ENDPOINT TOOL ACCESS

Featured

GitHub96.5K

thedotmack/claude-mem

Persistent Context Across Sessions for Every Agent – Captures everything your agent does during sessions, compresses it with AI, and injects relevant context back into future sessions. Works with Claude Code, OpenClaw, Codex, Gemini, Hermes, Copilot, OpenCode + More

youtube
Claude-Mem: Give Claude Code a Permanent Memory

Market Signal

Why It Has Market Pull

A genuinely popular, actively-discussed open-source memory layer for AI coding agents, with strong organic growth, a real Show HN launch, and detailed back-and-forth from skeptical and enthusiastic users alike. This is real builder momentum, not inflated numbers — the evidence base is unusually deep for a free, solo-maintained tool.

  • 96,581 real GitHub stars and 8,519 forks, confirmed directly against GitHub's own data
  • Tracked by a GitHub-trending aggregator with reported single-day growth of 627 new stars
  • Real Show HN launch with substantive discussion
  • Promoted by a known AI commentator on X, plus an independent third-party review blog
  • Spawned comparison/competition from at least one other project explicitly positioning against it
  • Supports Claude Code, OpenClaw, Codex, Gemini and other agent harnesses, not just one ecosystem

feedbacks

What People Are Saying

  • "I have tried probably 10-20 other open source projects... still nothing works better than simply keeping my own library of markdown files"HN comment

  • "I'd prefer not to send proprietary code to third-party servers"HN comment

  • "A claude.md file will give you 90% of what you need. Consider more when you're 50+ hours in"HN comment

  • "Is anyone else just completely overwhelmed with the number of things you need for claude code?"HN comment

  • "I didn't want a database or an MCP server or embeddings or auto-indexing when I can build something frictionless"HN comment

  • "You can now give infinite memory to Claude Code"X post

Product Hunt330

CoreSpeed

CoreSpeed is an operating system for agents. Connect apps once and bring your accounts and shared or private memory to Claude Code, Codex, Cursor and other MCP agents. Use multiple accounts per app, including X, with web search, social research and media generation built in. CoreSpeed holds app credentials. Budgets, activity logs and Smart Approval (beta) keep agents in check. Planned: Agent Drive, Mail, Pay, sandboxes, and more. Everything your agents need, through one MCP endpoint.

x
CoreSpeed — one MCP endpoint for your agents

Market Signal

Why It Has Market Pull

CoreSpeed looks like a real, well-funded company with genuine builder momentum, not just a one-day launch spike. It topped the daily leaderboard, has an open-source agent framework with hundreds of GitHub stars, and has already closed a multi-million-dollar seed round from named investors, making it one of the stronger agent-infrastructure plays among similar tools.

  • Finished #1 Day Rank on Product Hunt with 330 upvotes and 634 followers
  • Open-source Zypher agent framework (corespeed-io/zypher-agent) has roughly 961 GitHub stars
  • Raised a multi-million-dollar Seed++ round from Baidu Ventures and Monad Ventures, reaching a valuation in the tens of millions
  • Founders profiled by TechBullion and the company covered by China Daily for its agent-infrastructure work
  • Maker team actively answered feature questions (Slack/Discord connectors, per-agent access controls) live in the launch comments

feedbacks

What People Are Saying

  • "Slack and Discord are already available as connectors."Product Hunt comment

  • "pausing uncertain actions ... is the right call"Product Hunt comment

  • "activity logs and budgets"Product Hunt comment

  • "per-agent access restrictions ... isn't yet available but is planned"Product Hunt comment

  • "Gen Z innovators are shaping the future of AI"X post

  • "Building an Agent-Native Future and Monetising AI at Scale"Industry press

HF Spaces207 likes

OpenVuln

Find bugs in your repository with GLM This Space is built automatically from the root Dockerfile and serves the Vite application with nginx on port 7860. Optional Space build variables: VITEAPIBASEURL — API origin; defaults to https://openvuln.vulnhunter.pro. VITEGITHUBREPOURL — source repository linked from the interface. OpenVuln is a Hugging Face Space tagged with docker, region:us. It has 207 likes on Hugging Face.

Market Signal

Why It Has Market Pull

A production tool from a major AI lab, running on their flagship model to hunt for and responsibly disclose real vulnerabilities in open-source repositories. It has driven genuine independent scrutiny from major outlets and published safety research, making it the most credible and highest-impact entry in this batch.

  • 207 likes; backed by a major frontier AI lab, running on their flagship model
  • Real production usage: reported to have surfaced thousands of vulnerabilities across hundreds of open-source projects within weeks of its August 2026 launch, backed by a public disclosure ledger
  • High-profile independent scrutiny from major outlets directly engaged with the underlying model's cyber capability
  • Active Hacker News discussion debated both the capability and the safety framing, indicating real technical-community engagement rather than pure PR

feedbacks

What People Are Saying

  • "attackers can bypass the model's safeguards between 64 and 100 per cent of the time with simple techniques"Security research report

  • "open dangerous / closed safe"Hacker News comment

  • "Independent third-party coverage is still thin."evidence gap

Hacker News25 pts

OpenAPPA – open-source deterministic guardrails that don't break agents

Hi Hacker News! Matvey, one of the authors, is here. While building enterprise agents, we ran into a problem: the more tools you connect to the AI, the higher the chance it will run out of control and leak sensitive data. Guardrails, in theory, should prevent this, but the situation is worrying: - Non-deterministic guardrails (LLM as a judge, auto modes, etc.) are vulnerable to prompt injections, or they lack knowledge of the data, making them inefficient (~10% data leaks on our benchmarks). - E... (25 points, 12 comments).

youtube
Did a 50 year old military secret just solve agent prompt injection?

Market Signal

Why It Has Market Pull

A real, well-funded product with genuine builder momentum: backed by a security startup that closed a seed round led by a top-tier VC with angel investors including a well-known infrastructure-company CEO, and its GitHub repo has grown to over 1,400 stars in under two months. The HN launch itself was modest but is reinforced by independent trade-press coverage, an accepted academic paper, and a widely-viewed independent YouTube explainer.

  • GitHub repo shows 1,434 stars, 61 forks, 20 open issues, created roughly two months before research and still actively committed to
  • The company behind it raised a $10M seed round in mid-2026, total funding $13.5M, with angel investors including a well-known cloud-infrastructure company CEO
  • Show HN got 25 points / 12 comments; a separate related HN post about the same project got additional points/comments days later
  • Backed by a NeurIPS-accepted academic paper and covered independently by a trade publication reporting benchmark results
  • A popular independent tech-explainer YouTube channel (839K+ views) covered the project's core 'two-person rule' security model in depth

feedbacks

What People Are Saying

  • "My favorite part is 'batteries': you can run arbitrary programs as part of an authorization decision."HN comment

  • "Really interesting direction. What resonated with me is that you're treating agent security as an information-flow problem rather than a prompt-classification problem."HN comment

  • "Guardrails with builtin remediation instead of simply blocking my agent is a mind blowing long awaited experience!"HN comment

  • "i suspect we'll see more of this: flexible agents but deterministic boundaries. Congrats on launch!"HN comment

  • "Finally some determinism in our high-temperature sampling world!"HN comment

  • "the paper is good! thorough. I like it."HN comment

Hacker News13 pts

MCPJam - the first testing & evaluations platform for MCP servers

Prathmesh, CEO of MCPJam here. Users now start in ChatGPT, Claude, Cursor, and other AI clients. They reach your product through your MCP server. That means your users often aren’t in your product anymore. You can’t see what they prompted for, how the agent interpreted it, or whether your server helped them get the result they wanted. I saw this firsthand leading MCP technical strategy at Asana, including our ChatGPT and Claude launches. We were building high-stakes enterprise integrations, but... (13 points, 9 comments).

youtube
What is MCPJam Inspector | Exploit MCPJam Inspector

Market Signal

Why It Has Market Pull

MCPJam shows the most mature, sustained traction among similar tools evaluated here: over a year of GitHub history culminating in 2,200+ stars and nearly 300 forks, seed-stage VC backing, and a named enterprise customer with a published case study tying directly to the founder's claimed prior MCP background there. The specific HN thread evaluated here understates its real reach — it is one of at least ten related HN posts since mid-2025, and the official YouTube channel has published multiple hands-on walkthroughs.

  • GitHub repo shows 2,236 stars, 294 forks, 322 open issues, created roughly 17 months before research and still actively maintained
  • Show HN got 13 points / 9 comments; at least 9 other related HN submissions exist since mid-2025 under different framings
  • Backed by an early-stage venture fund per company profile data
  • Published customer case study describing a named enterprise customer building AI experiences with a tighter MCP feedback loop
  • Active product cadence post-launch, including a new enterprise platform layer shipped after the Show HN

feedbacks

What People Are Saying

  • "We have been a happy user of it! Excited to see the product add more capabilities!"HN comment

  • "how are you planning to stay compliant with the exploding number of clients users will have in practice?"HN comment

  • "yeah the number of clients will definitely increase, we'll start by looking to stay up to date with major AI clients as best we can"HN comment

  • "love the direction, but the problem for me has been about creating stronger evals and knowing what I should be checking for"HN comment

  • "Does this work with stdio MCP servers? Or just http/https ones?"HN comment

  • "test your mcp server, make sure it works effectively across clients :)"HN comment

Sources

GitHub

Panniantong/Agent-Reach

91.7K

Give your AI agent eyes to see the entire internet. Read & search Twitter, Reddit, YouTube, GitHub, Bilibili, XiaoHongShu — one CLI, zero API fees.

caddyserver/caddy

77.0K

Fast and extensible multi-platform HTTP/1-2-3 web server with automatic HTTPS

calesthio/OpenMontage

63.8K

World's first open-source, agentic video production system. 12 production pipelines, 100+ tools, 700+ agent skill and production-knowledge files. Turn your AI coding assistant into a full video production studio.

cloudflare/cloudflare-os

10.9K

Agent workspace built on Cloudflare Workers for creating documents, building apps, and running agents with your company’s context and systems.

Product Hunt

Gemini 4 Argon

292

Gemini 4 Argon is Google’s frontier AI model for complex, long-horizon professional work. It combines advanced reasoning, coding, multimodal understanding, and cybersecurity-defense capabilities, with support for up to 1 million output tokens. Designed for software engineering, finance, legal work, and enterprise research, it helps solve multi-step problems while Google gradually expands access through trusted testing and safety safeguards.

Smooth Recorder

232

Smooth Recorder is a native Mac app for demos and tutorials. It zooms in where you click, smooths your cursor, adds captions on your Mac and puts it all on a nice background. Screenshots get the same features.

Blume 2.0

202

Fast, AI-ready, and zero-config. Drop Markdown into a folder and ship a production-grade docs site with no app boilerplate to write or maintain. Free and open source, forever.

DocsAlot MCP Connector

173

Create and update your knowledge base from Claude, Codex, or any compatible MCP client. Ask DocsAlot to add a page, fix an error, or improve your help center. It finds the relevant content, makes changes, creates a new version, and publishes, all from your existing workflow. See what questions users ask in your help center, discover what’s missing, and use those insights to improve your docs and decide what to build next.

ChatGPT Space

159

ChatGPT Space is a shared canvas for creating, planning, and collaborating with AI. Add links, files, and ideas to build a living source of context, then work with ChatGPT to turn them into plans, docs, apps, and more - keeping your team aligned in one place.

opensend.cc

160

opensend.cc is an open source email platform that runs on your server. It sends through your own AWS account, so your domain, your data and your sender reputation stay yours. You get a REST API and SDK, SMTP, templates, broadcasts, automations, audiences, webhooks and logs. One command installs it. No per-email pricing, no feature gates. Already on Resend? Change two lines and keep your code.

YC Launch

Halmos Labs; intelligent experiment design

1

Halmos designs experiments for biotech teams Halmos Labs · Fall 2026 · Healthcare Tags: AI-powered Drug Discovery, Swarm AI, Biotech. Website: https://halmoslabs.com/

Hacker News

HN.watch – Videos of all Hacker News posts

225 pts

Hi HN, I’m Per, founder of Scrimba (YC S20). We’ve spent the last decade teaching people how to code with an HTML-based video format. We’ve now plugged an LLM into it, so that people can create explainer videos about anything. It’s called “Scrimba Explain”. To demo this technology for Hacker News, we built HN.watch. It’s like HN, but with explainer videos instead of articles. We create them on-the-fly the first time someone clicks on a link. While there are obvious visual drawbacks of using HTML... (225 points, 97 comments).

Made an open-source Lego AI generator

158 pts

Hi there :-) New on HN, first time posting. Past year, around December, I started experimenting with making ChatGPT and Claude generate source code in LDraw language. This LDraw is literally an "assembly" language, a low-level programming language that describes how to assemble LEGO pieces together into models, one placement instruction at a time. When executed by specific tools, like e.g. LDView, LeoCAD, Studio... these instructions become LEGO CAD models, that can be interacted with, modified,... (158 points, 50 comments).

PlaceCall (YC W26) – agentic API to call businesses and get things done

16 pts

Hi HN, this is Yarik and Vlad from VOYGR - we are building the tools for agents and apps to engage with local businesses. It all started with our own pain point at VOYGR: calling businesses to verify if they are open. We are both from Google (Maps and Search) and even there, the merchants and venues don’t keep this info updated. So we built an API and started using it in-house. On July 4th, we were driving through Portland looking for a place to eat. Google Maps was saying “Holiday hours may var... (16 points, 4 comments).

Praxos – Multiplayer AI

8 pts

Hey Hacker News! Lucas here, founder of Praxos (YC S24). Praxos is a team messaging platform for people and AI agents. It offers people and AI agents a place to talk and work together via a messaging platform that remembers the context around conversations. That context can then be used by the next person, AI agent, or even you, a week later. You can pick work back up without needing to get hold of another person to explain things again for you… or give you a refresher. A surprising amount of wo... (8 points, 0 comments).

Breadcrumb, record everything on your mac + context manager for AI

49 pts

Hi HN, I'm Justin. Breadcrumb records everything you do on your Mac (screen + meetings + AI transcripts + what you and your AI decided) and turns it into memory your AI can search. It's local and encrypted. You can also teach it rules by talking to it and it makes sure the right rules turn up in the right context. Works with Claude Code / Codex / Cursor / opencode. All of this is exposed to your AI as 30+ MCP tools (here's the definitions): https://innerloop.works/breadcrumb/mcp I started it in... (49 points, 9 comments).

Strata – an expressive semantic layer that can say no to your LLM

24 pts

Hello HN, I'm Ajo and I built Strata. I spent 4 years at Netflix solving self service for non-technical business users. I think I cracked it with my unique approach to semantic layer design. The key challenge is balancing expressiveness with ease of use for our non technical colleagues. It just so happens that focus made it work pretty well with LLMs too. Strata is a full stack solution. It includes a semantic layer, dashboards, subscriptions, and google sheets exports. All of it can be done vie... (24 points, 16 comments).

HF Spaces

Jev Decision Index

448 likes

Benchmarks and news on various repros of TypeSafe's Jev Who is rebuilding TypeSafe's Jev (System One / RLCD) in the open? This static Space opens on the Decision Index leaderboard; the News tab tracks the artifacts in one combined grid, color-coded by kind: Decoding: parallel constrained decoding on stock models (inference technique, no new weights) Diffusion: text diffusion models run in a "Jev mode" Trained: Jev-like scoring heads and fine-tunes, weights often on the Hub, promised models listed last Prior art: "this already exists" claims Explainers: architecture speculation, explainers, benchmarks and roundups Cards sort by a trending score: ♥ likes on X + 5 × GitHub stars + 8 × Hub likes...

Qwen-Image-2.1

377 likes

Interactive demo for Qwen-Image-2.1 — unified text-to-image generation and image editing with native RGBA transparency support. 📑 Blog 🤗 Model Weights 💻 GitHub Qwen-Image-2.1 is a Hugging Face Space tagged with gradio, region:us. It has 377 likes on Hugging Face.

Viggle Turbo for Qwen-Image-2.1

196 likes

6-step Qwen-Image-2.1, T2I + editing, vs-base comparison Viggle Turbo v0.3 — 6-step Qwen-Image-2.1 A distilled Qwen-Image-2.1 that generates and edits images in 6 steps with no classifier-free guidance, about 5× faster than the 40-step base model. On most prompts it is hard to tell apart from the base model; small, dense text and complicated edits (multi-reference composition, face swaps, identity-preserving edits) can still fall short of it. v0.3 (2026-09-29): at 6 steps, less grain than v0.2.1 and a little softer on fine texture. We think 6 steps is close to its capacity: every further gain we found cost something elsewhere. The new 9-step setting runs 7 turbo steps and lets the base model...

ChessFly

129 likes

Just a fruit fly's brain, playing chess Play chess against the complete connectome of an adult fruit fly. The FlyWire brain (138,639 neurons, 15.1M connections) runs in your browser on WebGPU: the board drives its 10,855 visual sensory neurons, activity settles over 5 steps across the whole graph, and the central brain and descending neurons are read out into a move and a win probability. Every forward pass lights up the 3D brain as it happens. The brain and the weights are about 150 MB, downloaded once and then cached. The wiring is fixed, exactly as the reconstruction has it, signs included. Only the encoder, the strength of each connection, each neuron's homeostatic gain and threshold, an...

MiniMax H3 Turbo LoRA

1.1K 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...

Wan2.2 14B Preview

320 likes

generate a video from an image with a text prompt Wan2.2 14B Preview is a Hugging Face Space tagged with gradio, mcp-server, region:us. It has 320 likes on Hugging Face.