A new Show HN submission can introduce a small, useful website long before it reaches a popular-products list. This edition collects ten recent submissions spanning technical practice, internal apps, spreadsheets, domains, browser privacy, games and testing tools.
Selection window: the 24 hours ending October 7, 2026 at 7:18 p.m. Pacific time (October 8, 02:18 UTC). We selected these from Show HN’s newest submissions and checked their product pages and discussion threads. This is an editorial selection, not a popularity ranking. A recent Show HN post does not necessarily mean the website itself was first created that day. Descriptions are based on public material rather than full product testing.
1. StayLeet: Keep technical judgment in practice
StayLeet offers two useful practice formats: short technical questions and longer system-design reviews. The site includes topics such as event loops, SQL and concurrency, alongside design exercises involving URL shorteners and webhook delivery. It is a relevant stop for engineers who want to keep evaluating systems even as agents write more code.
Before you try it: The focus is reasoning and review practice; it is not a substitute for building and operating a real system.
Visit website · Read the Show HN discussion
2. Railcode: Put internal apps behind shared controls
Railcode addresses what happens after employees build small apps and need to share them inside a company. It combines deployment with organization-level authentication, permissions, connectors and audit logs. Engineering teams dealing with a growing collection of internal tools may find the centralized management approach more interesting than another standalone app builder.
Before you try it: The maker describes support for static frontends and serverless JavaScript backends; check your runtime requirements.
Visit website · Read the Show HN discussion
3. SheetRelay: Give an existing spreadsheet an API
SheetRelay exposes Google Sheets through HTTP endpoints and MCP, with controls over which operations and data are available. A team could use it for a small content site, an internal form or an assistant that needs access to a maintained table. Its appeal is keeping the spreadsheet as the editing interface.
Before you try it: Review the exposed columns and write permissions before connecting a real sheet. Plans have monthly request allowances.
Visit website · Read the Show HN discussion
4. NetPanzer Browser Edition: Try a classic battlefield game in a browser
This Show HN submission brings NetPanzer into the browser through a WebAssembly build. The public play page explains how to join an existing server in Chrome or Firefox. It is a fun change of pace from developer tools and a concrete example of older native software finding a new distribution route.
Before you try it: A keyboard and mouse are required. The first load is substantial, and hosting a server still requires the desktop version.
Visit website · Read the Show HN discussion
5. Botscent: Inspect signals of agent-driven browsing
Botscent is a small library with a public demonstration site for identifying evidence of AI agents and crawlers. It combines request-level signals with page-level observations, returning reasons alongside its classification. Website builders can explore it for analytics or adapting an interface when an agent appears to be visiting.
Before you try it: Its documentation explicitly says a human result only means no agent evidence was found. Client-side detection is not proof of identity.
Visit website · Read the Show HN discussion
6. gtlds.fyi: Browse proposed domain extensions
gtlds.fyi provides a browsable view of proposed generic top-level domains and the companies applying for them. The interface includes length filters, views for overlapping applications and an option to hide brand names. It is particularly relevant to OpenURL readers researching naming ideas or following changes in the domain landscape.
Before you try it: An application is not an approved or available domain extension. Verify availability separately before planning a new address.
Visit website · Read the Show HN discussion
7. Buildra: Move from an app description to a running preview
Buildra lets a user describe an application and follow an agent’s work in a sandbox, then publish the result to an address. The site also describes downloading the source. It is a candidate for founders and makers who want to explore a small product idea before committing to a full development setup.
Before you try it: The service uses credits and project limits. Review generated behavior and deployment requirements before relying on the result.
Visit website · Read the Show HN discussion
8. Smart Blur: Reduce accidental exposure during screen sharing
Smart Blur is a browser extension with a product site and practice page for obscuring sensitive details. It combines pattern matching with manual blur controls; its paid features add local AI and screen-share detection. Presenters and support teams can explore it using the practice content before sharing real customer screens.
Before you try it: The maker notes that local AI needs WebGPU and a model download. It works on web pages, and automatic protection differs between web and desktop meeting apps.
Visit website · Read the Show HN discussion
9. getrequest: Retain API traffic for later investigation
getrequest places a request-handling layer in front of a backend. Its site describes retaining request details, inspecting traffic, replaying requests and retrying asynchronous delivery. Developers working with webhooks or unreliable downstream services can evaluate it as a way to make failures easier to inspect and recover from.
Before you try it: Some recovery features depend on the plan. Design replayed operations to avoid duplicate side effects, and review what request data is retained.
Visit website · Read the Show HN discussion
10. VoiceGremlin: Test a voice agent through an actual phone call
VoiceGremlin provides an API for testing AI phone agents with real calls and plain-language test goals. The service runs the conversation and returns a result and transcript, making it relevant to teams adding voice-agent checks to a release pipeline. This tests the call path as well as the conversation itself.
Before you try it: The result is evaluated by a model. Inspect transcripts and use a controlled test number when assessing important behavior.
Visit website · Read the Show HN discussion
A practical way to explore the list
Start with the problem closest to your work. StayLeet is a short learning stop; SheetRelay and Railcode address different ways to support internal tools; getrequest and VoiceGremlin concern operational checks. Keep the original discussions nearby: maker replies often reveal requirements that a homepage leaves out.
Today’s domain list
Here are the ten domains from this edition. The button loads their website links into OpenURL, replacing the tool’s current list. Review the queue, choose a batch size, then click Open batch when you are ready.
stayleet.com railcode.dev sheetrelay.com netpanzer.io botscent.nibnalin.me gtlds.fyi buildra.dev smartbuildlabs.com www.getrequest.io voicegremlin.com
Load these 10 websites in OpenURL
If the list does not load, copy the domains above and paste them into the bulk URL opener. For Smart Blur, the loaded link points to its product page rather than the company homepage.