How to use OpenClaw to build a Markdown web server
nexdrew used Claude for requirements and scaffolding, then OpenClaw agents connected to Slack and GitHub issues to build mkdnsite and a hosted service, with AI writing almost all code.
news.ycombinator.com
Suggested prompt
Suggested prompt written by Agents In Use, not the source author. Replace bracketed inputs and review tool permissions first.
Build [Markdown web server] from [requirements] with OpenClaw agents. Record features as [GitHub issues] and coordinate through [Slack team-lead agent]. Implement runtime HTML rendering and raw Markdown content negotiation, plus [search and MCP requirements], and prepare [release target].
Reach people exploring AI agents
See placements →Evidence at a glance
Sourced account- Source author
- nexdrew
- Original account
- Read the source ↗
- Published
- Source link checked
- Reported result
- The author reports v1.0.0 released March 16 and v1.4.1 March 28, with almost every line written by OpenClaw agents or Claude Code sessions.
- Human involvement
- The author supplied ideas and directed work through GitHub and Slack; the post does not describe code-review or testing steps.
A linked account supports attribution. It does not mean we independently reproduced its result.
Report an issue with this case
Open correction emailOpens your email app with the case link included. Nothing is sent automatically.
How the agent did it
- Refined the idea and requirements with Claude and generated a scaffold.
- Configured OpenClaw agents on a personal machine and connected them to Slack.
- Logged ideas as GitHub issues and assigned work through a team-lead agent.
nexdrew built mkdnsite after questioning a workflow that turns Markdown into HTML and then back into Markdown for agents. Bun's Markdown functionality suggested a server that could render HTML at runtime while serving the original Markdown to agents. The author first worked with Claude to refine the idea, establish requirements and scaffold the project.
After starting the project on March 7, they configured OpenClaw agents on their personal machine the following Friday and connected them to Slack. They then spent evenings and weekends recording ideas as GitHub issues and talking with a team-lead agent in Slack to pick up work and implement features. This process covered both mkdnsite and a hosted service. Individual Claude Code sessions also contributed code.
The author reports releasing mkdnsite version 1.0.0 on March 16 and version 1.4.1 on March 28. Almost every line of code was written by AI, either through autonomous OpenClaw agents or Claude Code sessions. The server supported runtime rendering, content negotiation, full-text search, MCP tools and several embedded rendering features, without a build step.
The post asks readers for feedback on the combination of features and its usefulness.
Who it's for: Developers exploring issue-driven agent implementation for Markdown publishing tools.
Prompt
Suggested prompt (written by Agents In Use, not the author's)
Source: Hacker News · nexdrew
Frequently asked questions
How did the author assign work to OpenClaw?
They logged ideas as GitHub issues and spoke with a team-lead agent in Slack to pick up work and implement features.
What role did Claude play?
Claude helped refine the idea, establish requirements and scaffold the project. Individual Claude Code sessions also wrote code alongside autonomous OpenClaw agents.
More details
Remaining steps
- Built mkdnsite and a hosted service through agent work and Claude Code sessions.
- Released mkdnsite v1.0.0 and subsequently v1.4.1.
About this account
Autonomous OpenClaw agents implemented features assigned through a Slack team lead, alongside individual Claude Code sessions.
Platform: Hacker News. Attribution: https://news.ycombinator.com/item?id=47606946. Prompt: Suggested (written by Agents In Use, not the author's). Sourced by the Agents In Use pipeline.
Credit: nexdrew on Hacker News
Source published
Added to Agents In Use
Source checked
Tools and topics
OpenClaw · Claude · Claude Code · Slack · GitHub · Bun
Markdown · Agent orchestration · Web server
Editorial completeness: 100/100. Measures documented fields, not agent performance or verified outcomes.