Issue tracking isn't dead. Your tracker just got bored of you.

Issue tracking is not dead. It is more important than ever, and the people who do it every day know that better than anyone. What got bored was one tracker, which decided the category was beneath its next act and pivoted to selling you agent management instead.

That is a real event, and it is worth being honest about. The best-built tracker of the AI era looked at the thing it does extremely well, a fast keyboard-first place to track the work, and called it finished. On Hacker News, the people who actually use it said the quiet part out loud: they just want a good issue tracker, and they are tired of watching every tool they like turn into something else.

We read those threads too. We built Radial for the person writing them.

#What "issue tracking is dead" actually means

It does not mean teams stopped tracking issues. Look at any repo, any sprint, any incident channel. The work still needs a home with structure: a status, an assignee, a priority, a due date, the relations between the thing that is blocked and the thing blocking it. That need did not shrink. If anything, more agents writing more code means more work to track, not less.

What the headline means is narrower and more commercial: a tracker decided the tracker is no longer the product, and the agent layer on top is where the new money is. That is a fine bet for a company to make. It is just not a bet that helps you if what you wanted was a tracker that opens fast and stays out of the way.

So the category is not dead. A position opened up in it.

#The Holdout's job

Someone has to keep building the boring one. The one that is still an issue tracker next year, and the year after. The one with no copilot to learn each quarter, no credit meter ticking in the corner, no quarterly relearning of where the buttons moved.

That is the whole pitch. An issue tracker. That's it.

• It opens fast and search returns before you finish typing. Speed is the feature, because latency is a focus tax and focus is the one thing a tool cannot give back to you.
• It runs from the keyboard. A command palette and real keyboard navigation, not a tour you dismiss.
• It is driven by your agent, not one we sell you. A first-class CLI, an MCP server, and a plain REST API. Claude Code, Codex, whatever ships next week, all drive Radial through real interfaces.

This is not anti-AI, and that distinction matters. AI is redefining a lot of products. Your issue tracker is not one of them. The intelligence belongs to your agent, your model, your keys. The tracker's job is to be a fast, trustworthy system of record that your agent can write to without you getting a bill for it.

#Verifiable today, not someday

The Holdout stance only counts if you can check it. So check it. Open a terminal and file an issue against your own workspace:

npm i -g radial.build
radial create "Stand up the new tracker" -t ENG -p high --json

That is the real CLI, doing the real thing, returning JSON your scripts and CI can read. Every command takes --json. Your agent can do the same over MCP at mcp.radial.build, or you can hit api.radial.build/v1 directly with a scoped, revocable key. The developer surface is shipped. Contrarian marketing is easy; shipping an OAuth server is the part most people skip.

And the part that keeps us honest is written down. Pricing is $50 per user, per year, flat, billed annually. Agents ride free, because every agent credential is a client of the API, not a billed seat. The rate you join at is the rate you keep. The pledge underneath it is binding: the day Radial ships a copilot, meters your usage, or charges you for AI you didn't ask for, your subscription is free. We are not asking you to trust a vibe. We wrote the consequence down.

#FAQ

#Is issue tracking actually dead?

No. One vendor declared its tracker finished and moved up the stack to agent management. Teams still track issues every day, and more code from more agents means more work to track, not less. The category did not die. A seat in it opened up.

#So is Radial anti-AI?

No. Radial is bring-your-own-agent. We never host, meter, or bolt in a copilot, but we ship the surface your agent runs on: a CLI, an MCP server, and a REST API. The intelligence is yours. The tracker is the system of record it writes to. The enemy is the uninvited copilot and the meter, never AI itself.

#Why does "your agents ride free" matter if other trackers also let agents in?

Agents getting access is becoming table stakes. The difference is the bill. In Radial an agent credential is a client, not a billed seat, and there is no AI credit meter anywhere in the product. One flat price, locked, with a written guarantee if we ever break it. The promise is what's unusual, not the API.

#What if Radial pivots too, or shuts down?

Fair question, and the reason we made two things commitments instead of vibes. Your data leaves in one command, JSON or CSV, always available, so you are never locked in. And "boring on purpose" is the roadmap, not an accident: Radial is built to still be an issue tracker next year, doing the same one thing well.

#How do I move off my current tracker?

Deep import from Linear or Jira brings your issues, projects, labels, comments, and history. Run it with --dry-run first to preview, then for real. The switch is meant to cost you a command, not a quarter.

#You wanted a tracker. Here's one.

If you read the threads and thought I just want a good issue tracker, this was built for you. Track issues like it's 2019. Ship like it's 2026.

See what we will and won't do in the manifesto, or read the one flat number on pricing. And on the calm-software thesis underneath all this, here is why being boring on purpose is the point.