Cloudron Forum

Just want to add to the appreciation here. Very well done, guys, thank you very much!

Not everyone is on a fixed IP all the time (residential addresses, mobile networks, etc.) … but why not both? Open a request

@girish Weird blog to be honest. Only two posts with the first one telling us LLMs will hack everything on the internet and the second one ranting about how LLMs didn't find an easy hack but deceived us. I am not skilled enough of doing any testing myself but I am thankful for your overtime hours and acting as fast as you did. If it happens to be not as bad as everyone feared, we can happily move on. As for me, I took it as an urgent reason to finally switch a legacy website. Working under pressure gets things done faster.

Hi, I’m trying to packaging an app that reverse-proxies uploads to a backend in real time (a tunnel/gateway app) to ease my setup. Cloudron’s nginx currently buffers the complete request body before passing it to the app (proxy_request_buffering on). For large uploads this means: clients see 100 % progress while the app hasn’t received any data yet, the whole body is written to the server disk first, the total upload time roughly doubles for streamed workloads. A manual test with proxy_request_buffering off; in the app’s vhost confirmed that streaming works as expected, but of course that change is overwritten whenever Cloudron regenerates the config. Would it be possible to add a manifest option (for example "proxyRequestBuffering": false), or an admin setting per app, to disable request buffering for apps that need streaming uploads? A related question: Cloudron replaces 502/503/504 responses from the app with its own appstatus.html page. Is there a way for an app to return its own error responses (for example JSON errors for API clients) instead? Your's Tim, Thanks!

[1.57.2] Update chatwoot to 4.19.1 Full Changelog Updated Node.js to the 24.22.1 security release. Hardened Twilio webhook verification and CSS sanitization. Added business hours and withheld caller ID support for Twilio calls. [Enterprise] Improved WhatsApp campaign recovery and duplicate-send protection. [Enterprise] Added SAML metadata URL support with certificate refresh. [Enterprise] Updated Captain AI and transcription models. [Enterprise] Added Stripe and auto-translation to paid Cloud plan entitlements. Fixed duplicate Help Center articles and conflicting inbox bots. Removed unnecessary session cookies from public pages. Thanks to @xamimran for the contributions.

@james good work. I thought I saw it on other packages, but maybe I am mistaken as I don't see others any more. I was just reviewing the list to check if any of my packages date from before eMudhra, and updating them out of caution (and moving to cloudron6 base images). All done - I dread to think what it's been like for you guys with so many more to do.

https://github.com/HarnessRouter/harnessrouter [image: HarnessRouter] The unified interface for agent harnesses. [image: GitHub Stars, exact count] [image: License: Apache 2.0] [image: Docker pulls] [image: UHP conformance: Full] [image: OpenAI Responses: Compatible] Build agent products without handling harness engineering. HarnessRouter is the infrastructure layer that turns existing harnesses like Codex and Claude Code into plug-and-play agent backends, letting your product run tasks, retrieve results, and switch harnesses through one API. [image: Help grow the HarnessRouter community. Star this repo →] [image: Animated diagram comparing separate harness integrations with one HarnessRouter API. Without HarnessRouter, four harnesses require 36 integration responsibilities, increasing to 45 with a fifth. With HarnessRouter, the product keeps one integration as harnesses are added.] [[image: Animated diagram comparing separate harness integrations with one HarnessRouter API. Without HarnessRouter, four harnesses require 36 integration responsibilities, increasing to 45 with a fifth. With HarnessRouter, the product keeps one integration as harnesses are added.] ](https://github.com/HarnessRouter/harnessrouter/blob/main/docs/images/2026-09-13-harnessrouter-integration-comparison-cropped-v5.gif) HarnessRouter implements the Unified Harness Protocol (UHP) and provides an OpenAI Responses-compatible API, handling persistent sessions, streaming progress, files, artifacts, cancellation, and structured failures. Switch harnesses. Optimize cost and latency. [image: Cost: 99.8% lower, from 223 to 0.47 credits. End-to-end latency: 3.2 times faster, from 4m 36s to 1m 25s. Eight harness and model configurations on the same task. Each metric compares its best and worst results. The lowest-cost and fastest configurations vary by task.] Eight harness × model configurations on the same task. Each metric compares its best and worst results. The lowest-cost and fastest configurations vary by task. Methodology 26 harness bases today: Codex, Claude Code, Hermes, Pi, DeepSeek Harness, OpenCode, Kilo Code, Qwen Code, Gemini CLI, Cline, Oh My Pi, goose, Kimi Code CLI, MiniMax Code, Aider, OpenHands, CheetahClaws, Grok Build, Agent Zero, System One, Qoder Cloud Agent, MiMo Code, jcode, OpenSquilla, nanobot and ZeroClaw. The support matrix shows which run on which provider. Tip Get started: Run locally · Integrate into your product · Starter kits · Use managed Cloud Quickstart Self-host Community Edition with your own provider keys, on infrastructure you control. Start with one Docker command, wait for the first launch, then connect a model provider and run your first task. You need: Docker · About 4 GB of disk · A provider API key No HarnessRouter account required. No bundled model or trial key.

Thanks, I have removed it.

I moved most of my apps to on-premise servers and kept a handful on a Hetzner VPS to keep costs down. I'm using SFF computers like the HP Elitedesk 800 G3 Mini (under $100 on eBay nowadays) and it's a solid machine for hosting. Older ram like DDR3/DDR4 are easier to find/buy too.