How billing works

You see the price after the bot scopes the work.

No menus to pick from. No tiers to choose. The bot reads what you're asking, quotes a flat number up front, and only starts when you approve.

First $50 of work is on us โ€” per workspace

Three steps. No surprises.

The bot is the one quoting. You're the one approving. We're the one eating any overrun.

STEP 01

You ask

Drop a task in Slack โ€” fix a bug, refactor, draft a PR, anything. Bot might ask clarifying questions if the scope is ambiguous.

STEP 02

Bot quotes

One flat number, up front. It tells you what's included and whether a human will be in the loop. You see the price before any work starts.

STEP 03

You approve

Say go, and work starts. Bill = quote, no matter what real cost is. If the bot under-quotes, that's our problem. Not yours.

What we guarantee.

Four promises baked into the model. Not marketing copy โ€” actual product behavior.

๐Ÿ‘‹

First $50 of work is on us

Per workspace. No card to start. Use it on whatever, however. See if the bot's worth it before you put money on the table.

๐ŸŽฏ

The quote is final

Bot says $X, you pay $X. Even if real compute or human time goes higher. We eat the overrun, you don't get a surprise bill.

๐Ÿชฅ

Honest routing

The bot tells you when human is needed โ€” and when it isn't. The recommendation isn't biased toward our margin. Sometimes it just says "I've got this."

๐Ÿ“ž

Unused credit refunds

If you uninstall, unused free quota or paid balance refunds. Per the privacy policy, your data wipes the same transaction.

Two ways your work gets done.

The bot decides which path fits โ€” execution work it handles itself, judgment work it routes to a human. You approve either way.

๐Ÿค– Most tasks

The bot handles it

Routine engineering work โ€” bug fix, refactor, scaffold a new project, write tests, draft a PR, explain code. Triggered by a Slack message, fulfilled async. No meeting, no booking, no waiting on a human's calendar.

Lighter, faster, the default path.
๐Ÿ‘ฅ When it matters

We add a person

For solution design, architecture choices, prod-touching changes, strategic calls โ€” the bot pulls you aside and recommends booking time with an experienced engineer. They lead it, write the deliverable, stay on call.

Heavier, considered, when stakes are real.

Common questions.

How do I know what something will cost before installing?

You don't, exactly โ€” and that's the point. Specific prices are quoted in-conversation, scoped to your actual task. Install with the $50 free credit, ask the bot to scope a real task you have, see what it quotes. No commitment.

What if the quote feels too high?

Push back in chat. "That's too much" or "make it cheaper" tells the bot to shrink the scope to fit a lower price. It'll adjust. You're never stuck with a number you didn't agree to.

What if a task runs over the quote?

We eat it. The number the bot quoted is the number you pay, full stop. If real compute or human time goes higher than expected, that's on us. We learn from it; you don't bear it.

Why is the free quota per workspace, not per user?

So a team of 20 doesn't get $1,000 of free agent time by virtue of headcount. The workspace is the billable unit; the free quota matches.

Does credit expire?

No. Top-up credit rolls over indefinitely. Uninstall any time and we refund unused balance.

Do I pay per seat?

No. miniCTO is paid per task, not per user. Everyone on the workspace can use it. You pay for the work the bot actually does for you.

Is there a monthly subscription?

Not yet โ€” usage patterns vary too much. Per-task matches what you actually use. If your team scales to "we want unlimited agent time and a dedicated engineer", email cz2440@columbia.edu and we'll talk enterprise.

Install and watch the bot scope your work.

$50 free per workspace. No card. The bot DMs you when it's in.

Add to Slack