I gave my coding agent an n8n/Zapier-style automation platform over MCP
Most MCP servers get used one tool call at a time, inside a chat. When the chat closes, the work stops. I wanted something else: tell my coding agent "every weekday at 9am, do X and email me" and have it build a real workflow that keeps running on its own . So I built WebArm24 . It's a hosted automation platform in the same space as n8n, Zapier and Make, except it's designed to be driven by an…
WebArm24 is a hosted automation platform, similar to n8n, Zapier, and Make, but it operates differently. Instead of being built on a canvas, it's driven by an agent over a Multi-Cloud Platform (MCP). The platform was developed by a user who wanted to automate tasks for their coding agent, such as fetching currency rates, summarizing them, and sending an email on weekdays.
To set up WebArm24, users create a prompt in Claude Code specifying the desired automation - in this case, a daily task of fetching USD rates for EUR, GBP, and JPY, summarizing them, and emailing the results. The agent then drafts the necessary pipeline, creates it, runs it once to check, and continues running on its own cron schedule without any open chat.
The platform connects to the MCP server using a simple command: "claude mcp add --transport http webarm24 https://webarm24.online/mcp --header Authorization: Bearer wak_...". New accounts receive free credits for use.
The server offers about 60 tools, categorized into Pipelines, the AI builder, Triggers, Storage, and Direct tools. Pipelines consist of steps that can use outputs from previous steps through templates. The platform supports various step types, including agent steps driven by LLMs, REST API calls, and JavaScript or Python code execution.
One of the standout features is the mcp step, which allows pipelines to call other MCP servers on a schedule, passing data between them. For instance, a sample pipeline calls TidyData MCP's public_holidays three times in parallel for US, UK, and DE holidays, merges the results, publishes them as a shared table, and emails the team the link—all within 5.3 seconds.
Building WebArm24 revealed that agents are better at editing workflows than drawing them, making failures more readable, and enabling agents to fix their own mistakes using tools like test step and debug chat. Each change is saved as a revision, and deleted pipelines can be restored. The platform also recommends using LLMs for build-time decisions while keeping run-time deterministic, and using clear input schema descriptions with examples to avoid agent misinterpretation.
WebArm24 operates on a pay-as-you-go credit system, with pipeline creation and editing free for users. Runs, AI builder calls, and direct browser or email tool calls consume credits, but new accounts start with free credits. The website (webarm24.online) and MCP endpoint (https://webarm24.online/mcp) provide installation docs, demos, and access to the official MCP registry (online.webarm24/webarm24).
Written by urgent.news from Dev.to's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
