Question 1

What does Runline actually do?

Accepted Answer

When an AI coding tool tries something dangerous on a laptop or build machine, Runline stops it before it runs. Your people keep working. Your security team gets a clear record of what was blocked and why.

Question 2

Who is Runline for?

Accepted Answer

Security leaders who need a real stop on risky AI actions, not another dashboard of advice. Engineering orgs rolling out Cursor, Claude Code, Codex, and similar tools across the fleet. Buyers who need a short path from demo to proof on real machines.

Question 3

How is this different from prompt rules or AI gateways?

Accepted Answer

Prompt rules ask the model to behave. Gateways watch traffic. Runline sits at the moment of action and can refuse the command. Pair it with the tools you already buy for discovery and logging; it does not replace them.

Question 4

Which AI tools does it cover?

Accepted Answer

The ones your engineers already use day to day: Cursor, Claude Code, Codex, and MCP-based agents, plus the common CLIs those tools call. Coverage expands with your fleet, not with a new IDE plugin for every team.

Question 5

Will this slow engineers down?

Accepted Answer

Human terminals stay normal. Only AI sessions hit the rules. Most teams say the first thing they notice is a blocked action that would have been a bad day, not a new workflow to learn.

Question 6

Do we have to change how people work?

Accepted Answer

No new IDE plugins. No prompt overhead. Engineers keep their tools and habits. Security owns the locked rules and the evidence trail.

Question 7

What kinds of actions get stopped?

Accepted Answer

Destructive and high-risk moves: wiping environments, pushing to the wrong place, changing production-adjacent systems, and similar commands that should never run unattended. You choose the bar. Runline enforces it.

Question 8

What does a POV look like?

Accepted Answer

A free 30-day evaluation on up to 25 endpoints. No card required. You install on a small set of engineering laptops, see real blocks and allow decisions, then decide whether to expand. Most teams start there before a quote.

Question 9

How fast can we get value?

Accepted Answer

Teams usually see meaningful signal in the first week of a POV: which AI tools are active, which risky actions show up, and whether the default rules match how your org works. A walkthrough takes about 30 minutes.

Question 10

How does pricing work?

Accepted Answer

Per endpoint, billed annually. POV is free for 30 days. Team and Enterprise are quoted on fleet size, SSO needs, and how you want alerts and evidence delivered. Start with a demo or a quote request if you already know the shape of the rollout.

Question 11

How do teams usually roll this out?

Accepted Answer

Start on engineering laptops, prove the stop works, then expand to CI runners and production-adjacent machines. Security owns the rules. Engineering keeps shipping. We size the quote to that path, not a big-bang cutover.

Question 12

How do we know it actually works?

Accepted Answer

Every allow or deny is recorded with who ran it. We also publish a public proof page: attack scenarios replayed against the policy engine on every commit, with results and honestly disclosed limits. Ask for the evidence pack during a POV.

Question 13

What if Runline is unreachable?

Accepted Answer

You choose the fail mode with your security team. The point of a POV is to see coverage, decision quality, and what happens when the control plane is offline, before you commit fleet-wide.

Question 14

What about compliance and procurement?

Accepted Answer

SOC 2 is in progress, with an evidence package available on request during enterprise evaluation. Tenant isolation, data residency options, and a subprocessors list are covered on our trust page and in the enterprise DPA.

Question 15

Can we answer the board when they ask what agents did?

Accepted Answer

Yes. Locked rules, who-ran-it records, and executive-ready briefings mean you are not scraping chat logs after an incident. That is usually the moment buyers stop shopping and start a POV.

Every question a buyer asks us, answered straight.

The product

Your engineers

Trying and buying

Proof and trust

The question we can't answer here: yours.

Every question a buyer asks us, answered straight.

The product

01What does Runline actually do?

02Who is Runline for?

03How is this different from prompt rules or AI gateways?

04Which AI tools does it cover?

Your engineers

01Will this slow engineers down?

02Do we have to change how people work?

03What kinds of actions get stopped?

Trying and buying

01What does a POV look like?

02How fast can we get value?

03How does pricing work?

04How do teams usually roll this out?

Proof and trust

01How do we know it actually works?

02What if Runline is unreachable?

03What about compliance and procurement?

04Can we answer the board when they ask what agents did?

The question we can't answer here: yours.