Skip to content

Pricing · all of it · USD

Pricing,
in public.

The guardrails are free. Our time costs money. Every number we charge is on this page.

The CLI is MIT and stays that way. Three services, each at a fixed price, for teams whose agents can reach production. Pro is planned — priced here before it exists, sold only once it does.

Bulkhead

PRICE LIST · EVERY ITEM



05Pro · planned$39/dev/mo

waitlist only · not charged · no date


  • GUARDRAILS BEHIND A PAYWALL0
  • EVIDENCE METERED0
  • PRICES NOT ON THIS PAGE0

USD · updated 2026-09-25

Price list: the CLI is $0, Policy Review $490, Agent Hardening $2,497, Watch $1,500 a month, Pro planned at $39 per active developer per month.

Free where it matters.

The CLI is the whole floor, not a trial of it. The rules, the completion check, the caps, the ledger and the receipts ship in the MIT package, and no future tier takes one back.

No. 01Shipped

Bulkhead CLI

$0

MIT · forever · 2 runtime dependencies

The whole floor. Every rule, check, cap and receipt.

Your copy
npx @bulkheadtools/cli init

Included


  • Dangerous calls denied by default26
  • More with --preset strict+6
  • Protected paths, incl. its own policy + hooks✓
  • Completion check: claims vs. evidence✓
  • Spend caps, per session / per day$5 / $20
  • Loop freeze✓
  • Hash-chained ledger + ledger verify✓
  • Receipts: md / txt / html, redacted✓
  • Uninstall: one command, keeps your policy + ledger✓
  • Telemetry (hooks use npx → npm registry)none

  • TOTAL$0.00
What it doesn’t do

Paid where it’s work.

Our time, at a fixed price, for teams whose agents can reach production credentials. Each button opens an email with the questions we’d ask anyway.

Review → Hardening → Watch · the $490 counts toward Hardening

  • No. 02USD · fixed

    Policy Review

    $490

    Once · async · 3 business days

    We read your stack against the incident corpus and hand back a policy diff you can merge.


    On payment

    • bulkhead.yaml diff, PR-ready, every rule commented1
    • Exposure memo, ranked by blast radius2 pp
    • Walkthrough call30 min

    Turnaround
    3 business days
    Credit
    All $490 toward Hardening, within 30 days
    Guarantee
    Fewer than 3 concrete fixes: no charge

    For

    A team that already runs Claude Code and wants a second read of what its agents can reach before something goes wrong.

    Not for

    Hands-on install and tuning across repos — that is Agent Hardening.

    We need

    • bulkhead.yaml and .claude/settings.json
    • Output of bulkhead report 10 and bulkhead inbox
    • Where your agents can reach production
    • The permission mode you run Claude Code in
  • No. 03USD · fixed

    Agent Hardening

    $2,497

    Fixed price · 48 hours

    We map every path from your agents to production, write the policy pack, install it and tune it on real traffic.


    On payment

    • Exposure map: every agent path to productionh 0–8
    • Custom policy pack + managed-settings snippeth 8–32
    • Verified install, tuned on real agent traffich 32–48
    • Team walkthrough45 min
    • Policy-tuning support by email30 days

    Turnaround
    48 hours
    Scope
    Up to 10 devs · 3 repos
    Price
    Fixed. No hourly meter

    For

    Startup teams whose agents can reach production credentials — database URLs, cloud tokens, CI deploy keys.

    Not for

    More than 10 developers or 3 repos in one pass. Tell us the size and we will say what fits.

    We need

    • Read access to the repos in scope
    • How your agents run: modes, overnight, CI
    • Where credentials live
    • Any incident or near-miss so far
  • No. 04USD · fixed

    Watch

    $1,500/mo

    Monthly · 3-month minimum

    We keep your policy current as Claude Code changes, and read the ledger with you when something happens.


    On payment

    • Policy re-tuning as Claude Code changesongoing
    • Incident forensics from the ledger≤ 1 bus. day
    • Re-auditquarterly
    • Pro, when it shipsincluded

    Minimum
    3 months
    Capacity
    5 teams at a time
    Price
    $1,500 a month, fixed

    For

    Teams that run agents unattended and want someone on the other end of the ledger.

    Not for

    A one-off check — that is a Policy Review.

Deliverables are real work on your stack: we run the corpus, read your bulkhead.yaml and bulkhead report, and write fixes as a diff you can review. We never publish a client’s name without their sign-off.

In the worksNo date yet

Bulkhead Pro

$39

/ active developer / mo
planned

No guardrail will ever move behind a paywall.

Pro will be the hosted part: evidence that outlives the laptop, and the team around it. You can’t buy it yet. The waitlist is an email, nothing is charged, and your answer decides what gets built first.

  • P.01Carbon CopyWill anchor your ledger head off-machine, closing the one gap a local hash chain has: a torn-off tail.
  • P.02Morning ReceiptWill send one overnight digest across every session, contradicted claims first.
  • P.03CountersignWill route risky actions to a second approver.
  • P.04Receipt BookWill host your receipts, with a view across the team.

Also planned: Pro free for public open-source repos. Watch clients get Pro included when it ships. Nothing on this band exists yet, and nothing on it is sold.

Three rules we price by.

  1. 01

    Never gate a guardrail.

    Every rule, check, cap and receipt in the CLI is MIT and free. Paid work adds our time — and, once Pro exists, hosting. Never a guardrail.

  2. 02

    Never meter evidence.

    Pro will bill active developers only — not tool calls, ledger entries or receipts. Keeping more evidence should never cost more.

  3. 03

    Publish every price.

    Everything we sell is priced on this page, including the thing we haven’t built yet.

Money
questions.

Anything else, ask hello@bulkhead.tools. A person replies.

Can we get a refund?

The Policy Review carries a written guarantee: if we find fewer than 3 concrete fixes, you aren’t charged. Agent Hardening: no refunds; the scope and price are fixed and agreed by email before you pay. Watch: no refunds; after the 3-month minimum you can cancel before any next month starts. Pro can’t be bought yet, so there is nothing to refund.

What counts as an active developer?

Planned: a developer whose agent sessions sent evidence to Pro in that billing month. Seats that sit idle wouldn’t be billed, and nothing would be charged per tool call, ledger entry or receipt. Pro doesn’t exist yet, so this is the plan, not a contract.

Can I buy Pro now?

No. It doesn’t exist yet, so we don’t take money for it and there is no date. The waitlist is an email with four boxes — which feature you’d pay for first — and those answers decide what gets built first.

Is there an open-source discount?

The CLI is MIT and free for everyone, open source or not. Pro is planned to be free for public open-source repos.

Does the Policy Review fee carry over?

Yes. The full $490 is credited toward Agent Hardening if you book it within 30 days of the Review.

What if we’re not a fit?

We’ll say so and point you at the free CLI. The guardrails don’t depend on paying us.