Epho is an API that allows running Claude Code, Codex or Opencode in a sandbox in the cloud. It abstracts away sandboxes, and allows running coding agents with a single HTTP request.
Epho came out of our own struggles with building our own AI analyst:\n- Sandboxes give you bare machines; you need to configure them for agentic workloads.\n- Each agent behaves differently, and you need to build integrations with each of them.\n- Sandbox providers are not very reliable, which means you need to figure out a multi-provider strategy to avoid failures.\n- Logging, artifacts, input/output, event streaming, and all of the other operational aspects need to be figured out.
We had to go through the pain ourselves. We got to a point where things got quite reliable, and it became more obvious to us that this should be a primitive on its own: send a POST request, get the events streaming back to you.
Epho is an agents-as-an-API product: you send a request, it spins up a sandbox, configures the chosen harness, clones your repos, and kicks off the agent. It takes care of automatic fallbacks across different providers, handles auth stuff, and just streams back the events and outputs.
It supports Claude Code, Codex and Opencode out of the box, and pretty much all the models they support out of the box. It streams the events back, handles attachments and output files, automatically manages the fallbacks on different sandbox providers, retries, and all the auth stuff. You just send a prompt, your repo, MCP servers you want to use with it, and it runs them.
I recorded a demo here to show a real example: https://youtu.be/HGfly1aytPA
I am quite excited for Epho, simply because I think it is a new primitive that would allow building agents into product a lot easier than it is today. We are running our agents on Epho on prod, so we'll keep maintaining it regardless, and we wanted to ship it as an independent product.
Epho is free to get started, and you can run it with Opencode's free models to get started with it.
I am quite curious to hear what you'd think and would love to get your feedback.
Cheers,\nBurak", "5": "2026-08-23T07:07:38.916785"} {"0": 5, "1": "hackernews", "2": "https://news.ycombinator.com/item?id=49363466", "3": "AWS new AZ in eu-west-2 - beware of unexpected results", "4": "AWS new AZ in eu-west-2 - beware of unexpected results. It seems as though AWS have brought a new AZ online in eu-west-2 (London) today with very little fanfare or indeed without telling their support staff. It appears as eu-west-2d. No AWS docs yet reflect this change.
Wanted to give anyone who uses this region a heads up just in case - at a client site it has caused a widespread bunch of issues as a result of Terraform failures etc. Depending on your codebase resources might be torn down and recreated unexpectedly. You can argue the case for whether this should cause you issues or not given this is the cloud etc but I just wanted to try and save someone else some hours of head scratching.
aws ec2 describe-availability-zones --region eu-west-2 --all-availability-zones --query 'AvailabilityZones[].[ZoneName,ZoneID,ZoneTYpe,State,OptInStatus,Messages]' \n[\n [\n "eu-west-2a",\n null,\n null,\n "available",\n "opt-in-not-required",\n []\n ],\n [\n "eu-west-2b",\n null,\n null,\n "available",\n "opt-in-not-required",\n []\n ],\n [\n "eu-west-2c",\n null,\n null,\n "available",\n "opt-in-not-required",\n []\n ],\n [\n "eu-west-2d",\n null,\n null,\n "available",\n "opt-in-not-required",\n []\n ],", "5": "2026-08-23T07:07:38.916857"} {"0": 6, "1": "hackernews", "2": "https://www.tomshardware.com/pc-components/microsd-cards/microsd-card-testing-database-celebrates-third-anniversary-with-133-petabytes-of-data-written-across-4-6-million-cycles-hundreds-of-cards-tested-to-failure-reveal-sandisk-as-the-outlier-with-6-failures-of-the-7-tested", "3": "MicroSD torture test: 133PB written, SanDisk fails 6 of 7 cards tested", "4": "MicroSD torture test: 133PB written, SanDisk fails 6 of 7 cards tested. ", "5": "2026-08-23T07:07:38.916879"} {"0": 7, "1": "hackernews", "2": "https://worksinprogress.co/issue/future-of-medicine/", "3": "Every disease is a policy failure", "4": "Every disease is a policy failure. ", "5": "2026-08-23T07:07:38.916889"} {"0": 8, "1": "hackernews", "2": "https://quaesitor.eu/silent-failures/", "3": "Giving LLM the schema docs made its wrong SQL answers more plausible, not rarer", "4": "Giving LLM the schema docs made its wrong SQL answers more plausible, not rarer. ", "5": "2026-08-23T07:07:38.916898"} {"0": 9, "1": "hackernews", "2": "https://sightlessscribbles.com/posts/a-thermodynamic-failure-of-the-soul/", "3": "A Thermodynamic Failure of the Soul", "4": "A Thermodynamic Failure of the Soul. ", "5": "2026-08-23T07:07:38.916906"} {"0": 10, "1": "hackernews", "2": "https://trpevski.com/blog/multi-agent-orchestration-routing-and-failure-recovery/", "3": "You shouldn't let agents talk to each other", "4": "You shouldn't let agents talk to each other. ", "5": "2026-08-23T07:07:38.916914"} {"0": 11, "1": "hackernews", "2": "https://github.com/harnessrouter/harnessrouter", "3": "Show HN: HarnessRouter: Unified interface for agent harnesses", "4": "Show HN: HarnessRouter: Unified interface for agent harnesses. Hey HN! We are building HarnessRouter, a canonical API for running Codex, Claude Code, Hermes, and other managed agent harnesses as your product backend.
Before building HarnessRouter, I used to build our own agent harness for our products. I tried LangGraph, agent SDKs from different vendors, pydantic, LLM tool use / function call, and so on. It's a very heavy lifting engineering effort, and I am disappointed about the agent deliveries compared to what Codex, CC can deliver. That changed my mindset. The frontier labs and famous open source communities are already putting so much engineering effort to build the world's best harnesses, why not leverage them directly instead of building our own, just like how we call LLM chat completion endpoints instead of training our own models?
We provide a docker image to run HarnessRouter locally.
----------
Quickstart:
docker pull harnessrouter/harnessrouter\n\n docker run -d --name harnessrouter -p 127.0.0.1:3000:3000 -v harnessrouter:/data harnessrouter/harnessrouter\n\n docker logs -f harnessrouter\n\n Wait for the "ready on :3000" show up, then open the browser at http://localhost:3000.\n Default username/password is harnessrouter/harnessrouter\n\n Then in Integrations page, add your model provider credentials or API keys.\n In Harnesses tab, as of today we provide routing to Codex, Claude Code, and Hermes as base harnesses.\n You can customize any of them and configure harness instruction, MCP tools, and skills.\n\n Then go to Tasks and let them do jobs.\nEvery harness has its own request/response format and incompatible with each other. We propose Unified Harness Procotol [1] to standardize how an application talks to an agent harness. It covers harness selection and configuration, task execution, event streaming, sessions start cancel and resume, artifact management and delivery, and failure handling. It's similar idea like LiteLLM, but for harnesses rather than models.
HarnessRouter implements UHP. We provide an AGENTS.md [2] and your coding agent can follow it to integrate your application with the harnesses available.
We also provide starter kits [3] to demonstrate some types of agentic products that can be built on HarnessRouter. It currently includes PPT agent, Spreadsheet agent, BI Dashboard agent, and Video generation agent.
Can't wait to hear what you think!
[1] https://unifiedharnessprotocol.org
[2] https://harnessrouter.ai/agents.md
[3] https://github.com/harnessrouter/starter-kit", "5": "2026-08-23T07:07:38.916923"} {"0": 13, "1": "hackernews", "2": "https://julienreszka.com/blog/failure-takes-years-to-admit/", "3": "Failure Takes Years to Admit", "4": "Failure Takes Years to Admit. ", "5": "2026-08-23T07:07:38.916941"} {"0": 14, "1": "hackernews", "2": "https://www.myrmikan.com/pub/Myrmikan_Research_2026_08_14.pdf", "3": "AI Debt Failure Will Prompt Another Wave of Fed Bailouts [pdf]", "4": "AI Debt Failure Will Prompt Another Wave of Fed Bailouts [pdf]. ", "5": "2026-08-23T07:07:38.916950"} {"0": 15, "1": "hackernews", "2": "https://news.ycombinator.com/item?id=49321344", "3": "Show HN: ViaNJ \u2013 an app that makes NJ Transit ticketing less painful", "4": "Show HN: ViaNJ \u2013 an app that makes NJ Transit ticketing less painful. I\u2019m building ViaNJ because I want a more seamless experience with NJ Transit departures, ticket purchase, activation, and getting a barcode ready while boarding.
It has live departures, alerts, track predictions, ticket purchase, automatic ticket recovery, Apple Wallet, and iCloud backup. The interesting technical part is making NJ Transit\u2019s unofficial web ticket flow reliable: ViaNJ keeps purchases recoverable through redirects/failures and only considers checkout complete once the actual ticket is recovered.
It\u2019s SwiftUI with a small backend and Cloudflare services. NJ Transit still processes all ticket payments directly, and ViaNJ never stores card data.
Still actively working on it, and accept any and all feedback. Fun fact, NJT is apparently a working on integrating Apple Wallet ticketing into their existing app. Unfortunately, I don\u2019t see their app UX improving much more, but happy to be proven wrong eventually.
https://www.vianj.app\nhttps://feedback.vianj.app/", "5": "2026-08-23T07:07:38.916959"}