Agno

Agno AI

Agno AI

Agno AI

Agno AI

Agno AI

Agno AI

Agno AI

Agno AI

Agno AI

Agno AI

Show HN: Coasts – Containerized Hosts for Agents

Hi HN - We&#x27;ve been working on Coasts (“containerized hosts”) to make it so you can run multiple localhost instances, and multiple docker-compose runtimes, across git worktrees on the same computer. Here’s a demo: <a href="https:&#x2F;&#x2F;www.youtube.com&#x2F;watch?v=yRiySdGQZZA" rel="nofollow">https:&#x2F;&#x2F;www.youtube.com&#x2F;watch?v=yRiySdGQZZA</a>. There are also videos in our docs that give a good conceptual overview: <a href="https:&#x2F;&#x2F;coasts.dev&#x2F;docs&#x2F;learn-coasts-videos">https:&#x2F;&#x2F;coasts.dev&#x2F;docs&#x2F;learn-coasts-videos</a>.<p>Agents can make code changes in different worktrees in isolation, but it&#x27;s hard for them to test their changes without multiple localhost runtimes that are isolated and scoped to those worktrees as well. You can do it up to a point with port hacking tricks, but it becomes impractical when you have a complex docker-compose with many services and multiple volumes.<p>We started playing with Codex and Conductor in the beginning of this year and had to come up with a bunch of hacky workarounds to give the agents access to isolated runtimes. After bastardizing our own docker-compose setup, we came up with Coasts as a way for agents to have their own runtimes without having to change your original docker-compose.<p>A containerized host (from now on we’ll just say “coast” for short) is a representation of your project&#x27;s runtime, like a devcontainer but without the IDE stuff—it’s just focused on the runtime. You create a Coastfile at your project root and usually point to your project&#x27;s docker-compose from there. When you run `coast build` next to the Coastfile you will get a build (essentially a docker image) that can be used to spin up multiple Docker-in-Docker runtimes of your project.<p>Once you have a coast running, you can then do things like assign it to a worktree, with `coast assign dev-1 -w worktree-1`. The coast will then point at the worktree-1 root.<p>Under the hood the host project root and any external worktree directories are Docker-bind-mounted into the container at creation time but the &#x2F;workspace dir, where we run the services of the coast from, is a separate Linux bind mount that we create inside the running container. When switching worktrees we basically just do umount -l &#x2F;workspace, mount --bind &lt;path_to_worktree_root&gt;, mount --make-rshared &#x2F;workspace inside of the running coast. The rshared flag sets up mount propagation so that when we remount &#x2F;workspace, the change flows down to the inner Docker daemon&#x27;s containers.<p>The main idea is that the agents can continue to work host-side but then run exec commands against a specific coast instance if they need to test runtime changes or access runtime logs. This makes it so that we are harness agnostic and create interoperability around any agent or agent harness that runs host-side.<p>Each coast comes with its own set of dynamic ports: you define the ports you wish to expose back to the host machine in the Coastfile. You&#x27;re also able to &quot;checkout&quot; a coast. When you do that, socat binds the canonical ports of your coast (e.g. web 3000, db 5432) to the host machine. This is useful if you have hard coded ports in your project or need to do something like test webhooks.<p>In your Coastfile you point to all the locations on your host-machine where you store your worktrees for your project (e.g. ~&#x2F;.codex&#x2F;worktrees). When an agent runs `coast lookup` from a host-side worktree directory, it is able to find the name of the coast instance it is running on, so it can do things like call `coast exec dev-1 make tests`. If your agent needs to do things like test with Playwright it can so that host-side by using the dynamic port of your frontend.<p>You can also configure volume topologies, omit services and volumes that your agent doesn&#x27;t need, as well as share certain services host-side so you don&#x27;t add overhead to each coast instance. You can also do things like define strategies for how each service should behave after a worktree assignment change (e.g. none, hot, restart, rebuild). This helps you optimize switching worktrees so you don&#x27;t have to do a whole docker-compose down and up cycle every time.<p>We&#x27;d love to answer any questions and get your feedback!

Spending too much time on content creation? Writesonic accelerates your workflow with AI. Biosynth scaled scientific product descriptions from 250 to 5,000 per week using Writesonic. 🌿 Discover the w

Spending too much time on content creation? Writesonic accelerates your workflow with AI. Biosynth scaled scientific product descriptions from 250 to 5,000 per week using Writesonic. 🌿 Discover the wisdom of your body. Free diagnosis from the profile link. #AItools #ContentCreation

Poison at play: Unsafe lead levels found in half of New Orleans playgrounds

Sarah Hess started taking her toddler, Josie, to New Orleans’ Mickey Markey Playground in 2010 because she thought it would be a safe place to play after Josie had been diagnosed with lead poisoning. Hess had traced the problem to the crumbling paint in her family’s century-old home. While it underwent lead remediation, the family stayed in a newer, lead-free house in the Bywater neighborhood near Markey, where Josie regularly played on the swings and slides. “Everyone was telling us the safest place to play was outside at playgrounds, so that’s where we went,” Hess said. Josie’s next blood test was a shock. “It skyrocketed,” Hess said. Josie’s lead levels had leapt to nearly five times the national health standard. When the soil at Markey was tested in late 2010, it too was found to have dangerously high levels of lead. But the city took no meaningful action to inform Markey’s users or make the park safe. Parents started posting warning signs at the park and flooded City Hall with outraged calls and emails. Holding Josie in her arms, Hess made an impassioned speech to the City Council.  A child’s shoes are left in the dirt next to the playground at Mickey Markey Park in the Bywater neighborhood of New Orleans in November 2025. It’s common for children to play barefoot at this playground. Christiana Botic / Verite News and Catchlight Local / Report for America In short order, the city had hired a company to test Markey and other parks, and pledged to fix the lead problem wherever it was found. “I couldn’t have been more pleased,” Hess said. “They were totally into it. My impression was they were going to make them all lead-free parks.” But a Verite News investigation conducted over four months in 2025 found that lead pollution in New Orleans parks not only persists, it is more widespread than previously known. Dozens of city parks with playgrounds remain unsafe, including Markey and others that underwent city-sponsored lead remediation in 2011. The city does not appear to have conducted any major remediation or lead testing of parks since that time. The findings indicate that city officials fell short in their cleanup efforts then, and that a very large number of New Orleans children are exposed to excessive amounts of lead now, said Howard Mielke, a retired Tulane University toxicologist and one of the nation’s leading experts on lead contamination. “It’s a failed program,” he said. “They didn’t do what they needed to do to bring the lead levels down in a single park.” Verite News reporters tested hundreds of soil samples from 84 city parks with playgrounds in fall 2025. Adrienne Katner, a lead contamination researcher with Louisiana State University, verified the results. The testing found that about half the parks had lead concentrations that exceed [a federal hazard level](https://www.epa.gov/newsreleases/biden-harris-administration-strengthens-safeguards-protect-families-and-children-lead) established in 2024 for soil in urban areas. “I am surprised they haven’t been tested and mitigated,” said Gabriel Filippelli, an Indiana University biochemist who studies lead exposure. “If there’s evidence of kids playing in soils that are as high as [Verite’s testing] described, that’s kind of horrifying.” Public health researchers and doctors say that children under 6 absorb lead-laden dust more easily than adults, contaminating their blood and harming the long-term development of their brains and nervous systems. There is no known safe exposure level for children, and even trace amounts can result in behavioral problems and lower cognitive abilities. ### **Find out what the lead levels are at New Orleans playgrounds** New Orleans is in financial straits with a [budget deficit](https://veritenews.org/2026/01/07/helena-moreno-interview-mayor/) of about $220 million, and it’s unclear what priority or resources Mayor Helena Moreno will, or even can, allocate to restart lead remediation efforts. In response to the financial crisis, Moreno has eliminated dozens of positions and plans to [furlough 700 employees](https://veritenews.org/2026/01/27/new-orleans-moreno-cuts-layoffs-deficit/) one day per pay period to save money. Moreno’s administration did not respond to requests for comment. The city doesn’t routinely test for lead in parks, said Larry Barabino, chief executive officer of the New Orleans Recreation Development, or NORD, Commission, the agency that oversees most of the city’s parklands. He confirmed the last significant effort to test parks ended in 2011. He called Verite’s results “definitely concerning” and pledged to work with city departments and local experts to potentially remediate unsafe parks. “Safety is our number one priority here at NORD,” Barabino said. “If there’s anythin

I just got ChatGPT Pro - is it really worth $200 a month 🤔 Will I be sad when I downgrade back to $20 a month Plus? No and no. But let's dive into why and how Pro is different. There are basically fo

I just got ChatGPT Pro - is it really worth $200 a month 🤔 Will I be sad when I downgrade back to $20 a month Plus? No and no. But let's dive into why and how Pro is different. There are basically four headline features of ChatGPT Pro: 1. Sora, the video generator. 2. Operator, the autonomous agent that uses the internet for you. 3. Deep Research, which is an agent-like researcher that goes off and does very in-depth research for you and returns very in-depth, long reports. 4. o1 Pro Mode, which is their most advanced reasoning model, which is great for very specific things like scientific research, financial modeling, medical diagnosis, science and maths. And coding as well, which are very specific applications. The other thing that Pro has, people don't talk so much about, is a much longer context window. If you're paying $20 a month for ChatGPT Plus, your context window, or conversation length, is only about 25,000 words. Which might sound like a lot, but if you're going back and forth with it, exchanging information, feeding back, inputting documents, you can eat up that 25,000 words quite quickly. Particularly if you start coding, that melts away. ChatGPT Pro has a context window of about 100k words and I think I’ve felt the benefit of that. Compare that with the fact that Claude and Gemini have absolutely massive context windows. Google Gemini Pro has a context window of 1.5 million words. So that alone isn’t worth $200 a month. What about those key features? So those four headline features is what really attracts people. But I'll be waiting for them to mature and make their way down to the Plus plan! #chatgpt #aitools #ai #artificialintelligence #learnontikok #tech #technology #openai #GPT #aiagent