writing-humanizer
@randalolson/writing-humanizerSafety
Runnable
Well-formed
Remove signs of AI-generated writing from prose text. Use when editing a draft that sounds AI-generated, reviewing content before publishing, cleaning up LLM output for human readers, or preparing text for a brand voice that requires natural human writing.
Lift the share of edited drafts that clear an AI-detection or human sniff-test review and reach publication, so readers engage with the prose instead of dismissing it as AI-generated.
See it in action
An AI wrote the first version of this launch post. The product is real and the details are good. They are just buried under a pile of annoying AI writing tropes.
What the AI wrote
We're beyond excited to introduce Field, a smart weather station that's set to revolutionize the way you connect with your own backyard. In an era of growing climate uncertainty, hyperlocal data has never been more crucial.
Field isn't just another gadget—it's a testament to what's possible when cutting-edge AI meets thoughtful design. Packing research-grade sensors into a palm-sized enclosure, it tracks temperature, humidity, soil moisture, wind, and pressure every sixty seconds, empowering you to truly understand the microclimate right outside your door. What began as a humble Raspberry Pi on a fence post has grown into something remarkable—a community of over 4,000 backers who raised $1.2 million on Kickstarter. Industry experts agree: the future of gardening is hyperlocal, and this is only the beginning.
The tells are the usual suspects: "beyond excited," "set to revolutionize," "a testament to what's possible," "industry experts agree," "this is only the beginning." Em dashes everywhere.
After the workflow
This is Field, a weather station small enough to sit in your palm. Inside it are sensors for temperature, humidity, soil moisture, wind, and pressure, and it takes a reading from all of them every sixty seconds. The idea is simple: the weather right outside your door is its own little climate, and a forecast for the whole county does not really tell you about it.
It started as a Raspberry Pi I taped to a fence post. I half expected that to be the end of it. Instead, more than 4,000 people backed it on Kickstarter and put in $1.2 million. I am still a little stunned that the fence-post version turned into this.
Same facts in the rewrite: the five sensors, the palm-sized build, the reading every sixty seconds, the fence-post Raspberry Pi, the 4,000 backers and $1.2 million. The difference is it has a voice now that doesn't sound like every other AI-written post.
What it changed
Some lines kept their facts and lost the gloss. The pure slogans got cut. Nothing was invented.
The receipt
The template scores its own output against the same nine criteria every run uses.
This is the same verifier the workflow runs on every output, not a one-off for the demo. A draft does not ship until all nine criteria come back clean.
Use this template
Your agent fetches this runbook and runs it, revising the output until the verifiers pass.
First time with Goodeye?
Connect your agent over MCP, then ask it to fetch this template by its identifier. Or install the CLI and run the command below.
claude mcp add --transport http goodeye https://mcp.goodeye.dev/mcpuv tool install goodeyeConnecting over MCP prompts a quick sign-in. The CLI can fetch and run a public template with no account.
Fetch or fork
goodeye templates get @randalolson/writing-humanizergoodeye templates fork @randalolson/writing-humanizerAlso available from
get_template(identifier="@randalolson/writing-humanizer")curl https://api.goodeye.dev/v1/templates/@randalolson/writing-humanizer