The short version

We have shut down Supp, the product. The dashboard, the chat widget, the API, the MCP server, and the SDK are offline, and account pages now show a notice instead of a login form.

Supp never had a paying customer. That makes this announcement simpler than most. Nobody loses data they depend on, nobody has an invoice to dispute, and nobody needs a migration guide. If you signed up to try it, your account no longer works and there is nothing you need to do.

What stays: this blog and the free tools. Every article is still here at the same address. The calculators, templates, and generators still work, still cost nothing, and still need no signup.

What Supp was

Supp was a support classification and routing tool for small teams. A message came in through a chat widget or the API, a classifier we trained ourselves put it into one of 315 intents, and Supp either answered it, ran an action, or sent it to the right person in Slack, Linear, Jira, or wherever the team already worked. Pricing was per resolved message, with no seats and no monthly fee.

We built it because we thought small teams were overpaying for help desks they barely used. Most of their messages were the same handful of questions, and a full help desk with per-agent pricing felt like the wrong tool for that.

Why we are stopping

We did not get people to pay for it. We shipped the product and wrote a lot about support, but reading about a problem did not turn into paying for our fix.

We could have kept it running and kept hoping. Running a classifier, an API, and a database for zero customers costs real money and attention every month. We would rather spend that on things people are already using.

What happens to each part of the site

The blog. Stays up. Same URLs, same content. We will keep correcting facts when vendor pricing or features change, and posts that get a material update show the date they changed.

The free tools. The calculators and template generators stay up. They run in your browser, so nothing you type goes to a server. That includes the cost, ROI, staffing, SLA, NPS, CSAT, CES, churn, and lifetime value calculators, the handbook and SLA templates, and the macro, apology, and response generators.

The AI tools. The live classifier playground and the support readiness grader are retired. Both called our backend, and that backend is gone. Their old URLs now point to the tools page.

The product pages. Pricing, docs, the dashboard, signup, and login are gone. Old links land on a short notice that points back here.

The comparisons. The head-to-head guides and the alternatives guides for Zendesk, Intercom, Freshdesk, Gorgias, and Help Scout stay up. We removed the parts that pitched Supp, since there is nothing left to pitch.

What we learned

We are not going to dress this up as a list of hard-won wisdom. A few things were plain enough that we would act on them next time.

Interest in a topic is not demand for a product. People read articles about support costs because support costs are a real problem. That did not mean they wanted to change tools to fix it. We treated interest in the writing as a signal for the product, and it was not one.

Small teams often do not feel the pain we were solving. A team that gets a few dozen messages a day can sort them by hand in a few minutes. The cost of adding a new tool, learning it, and trusting it was higher than the cost of the problem for a lot of the teams we were building for.

Start with the small, free thing. The calculators were quick to build and asked nothing of the people using them. If we had started there and asked users what they still struggled with, we would have learned more, sooner.

Shut things down while it is still cheap. With no customers, closing the product is mostly a matter of turning servers off. It would have been much harder with contracts and data to hand back.

Where to find us

Right here. The blog is now the home page, and new posts and corrections will show up there first. If you use a feed reader, the RSS feed at /feed.xml and the JSON feed at /feed.json still work.

If you used any of the free tools to plan a budget, a hire, or an SLA, thank you. That part of Supp was worth building, and it is not going anywhere.