Two numbers, no hourly.
Sprint Zero is a fixed-price pilot, $3.5k to $7.5k depending on scope, and it credits toward a retainer if you continue. Retainers run $5k to $15k a month for a standing slice of the machine. Both numbers are on this page because we mean them.
Rescue is quoted after we read the code, never before. No hourly billing, because hours are the wrong unit for a studio that runs agents all night.
- hourly
- none, on anything
- credit
- Sprint Zero's price comes off the first month of a retainer
- quoted
- means we read the code or the workflow before we name a number
- reply
- an engineer, within one business day, Pacific time
The scanner
Find out whether your AI-built app would survive a real user, a real attacker, and a real Monday.
The teardown call
Twenty minutes with an engineer, not a closer, on your actual repo, workflow, or plan.
The 7-day build
Give us one narrow problem. In seven days you get a working slice, built by the fleet and reviewed by a person, with the build log public.
The automation map
Thirty minutes. One page. Which of your workflows an agent should own first, and why.
Sprint Zero
A spec you can sign, a working slice, an eval harness that defines done, and a fixed price for the rest.
- Discovery, then a written spec with explicit non-goals.
- A working slice deployed, not a slide.
- An eval harness that defines done in a way a machine can check.
- An architecture and cost model, including token cost per task.
- A fixed-price quote for the load that follows.
- for
- SMB owners with a mapped workflow, founders with a proof build or a roadmap.
Rescue
Your AI-built app made boring: secure, tested, monitored, and ready for real users. We keep what works and rebuild only what does not.
First agent
One workflow, owned by an agent, with a human approval gate and a dashboard your bookkeeper trusts.
Product build
A production-grade product in four to six weeks: the fleet does the volume, a senior engineer owns the architecture, and you own everything.
Knowledge system
Every document, ticket, and contract your company has, answerable with the permissions you already have.
Capacity
A standing slice of the machine every month: a named engineer, a fleet budget stated in dollars, and a ship log you can read.
- A named senior engineer who owns your roadmap and reads every diff.
- A monthly fleet compute budget, stated in dollars in the proposal.
- A Monday plan, daily commits you can see, a Friday demo, and a ship log.
- Rollover of unused capacity for one month.
- The house standard on everything shipped.
- for
- Founders after a rescue or a product build, and companies after their first agent.
Free doors first. Sprint Zero credits toward a retainer if you continue. Everything else is priced as its own load. Quoted means we read the code or the workflow before we name a number.
Start with something free.
Four doors, one for each kind of visitor. Each one gives you something you can use whether or not you ever pay us.
No countdowns and no sales sequence behind any of them. The 7-day build is the one with a cap, a few a month, because it is gated by application.
The scanner
Find out whether your AI-built app would survive a real user, a real attacker, and a real Monday.
- A read-only scan of the URL you paste, with findings ranked by how much they can hurt you.
- Nothing stored unless you ask us to send the result to your email.
The teardown call
Twenty minutes with an engineer, not a closer, on your actual repo, workflow, or plan.
- What is fine, what is dangerous, and what to do first, in writing after the call.
- If it fits an offer on this page, we say which one and what it would cost. If it does not, we say that.
The 7-day build
Give us one narrow problem. In seven days you get a working slice, built by the fleet and reviewed by a person, with the build log public.
- One workflow, one integration, or one screen, deployed where you can use it.
- The fleet's task log, test results, and cost per task on a public build page.
- A follow-up call where we tell you what we would build next and what it would cost.
The automation map
Thirty minutes. One page. Which of your workflows an agent should own first, and why.
- A one-page map of the workflows we can see from the call: what runs on a spreadsheet, a shared inbox, or a person's memory.
- A first-agent recommendation with the human approval gate drawn in.
$3.5k to $7.5k, fixed.
A spec you can sign, a working slice, an eval harness that defines done, and a fixed price for the rest.
What moves the price inside the range: how many systems the slice has to touch, whether it needs real integrations or can stub them, and whether there is a spec already or we are writing it from a conversation. We tell you which end you are at before you sign.
If you continue to a retainer, the sprint price credits against it. If you stop, you keep the spec, the slice, the eval harness, and the quote.
Fixed price, $3.5k to $7.5k. Credits to a retainer if you continue.
- Discovery, then a written spec with explicit non-goals.
- A working slice deployed, not a slide.
- An eval harness that defines done in a way a machine can check.
- An architecture and cost model, including token cost per task.
- A fixed-price quote for the load that follows.
- systems
- how many the slice has to touch
- integrations
- real, or stubbed for the sprint
- spec
- exists already, or written from a conversation
- low end
- $3,500: one system, stubs allowed, a spec in hand
- high end
- $7,500: several systems, real integrations, spec from scratch
- for
- SMB owners with a mapped workflow, founders with a proof build or a roadmap.
- credit
- The full sprint price, against the first month of the retainer if you continue.
- if you stop
- You keep the spec, the slice, the eval harness, and the quote.
- rollover
- Unused retainer capacity rolls forward one month, not forever.
One load, one price, one shipping date.
A load is one scoped unit of work with a definition of done. Each of these is quoted after we have read the code or the workflow, because a number typed before that is a guess with a dollar sign on it.
Each ships with the house standard.
- A ranked list of what is wrong, in order of how much it can hurt you, before you pay for anything.
- Auth and data-layer fixes, migrations that run twice without damage, backups that restore.
- Tests on the paths that matter, so it stays fixed.
- CI, observability, cost controls, and a handover document.
- The option to move to a retainer, stated plainly in the handover.
- An intake, support, scheduling, invoicing, or phone agent wired into the systems you already use: CRM, accounting, phone, email.
- Bounded autonomy: approval thresholds written down, and a person on anything irreversible.
- An audit trail of every action the agent took and why.
- Evals that run before every change, and a dashboard with a cost line.
- Tuning after it goes live, because the first month teaches everyone something.
- Product spec and a design system you can extend.
- A full-stack build with evals and CI, deployed on your infrastructure.
- Analytics, a security baseline, and the house standard: evals, observability, a runbook, a 30-day fix warranty.
- Then, if you want it, a standing slice of the machine to keep shipping.
- Permission-aware retrieval over the stores you have: SharePoint, Google Drive, email, ticketing.
- A citation on every answer and a log of the questions it could not answer.
- Evals that test retrieval separately from generation.
$5k to $15k a month.
A standing slice of the machine every month: a named engineer, a fleet budget stated in dollars, and a ship log you can read.
The band is a band because the slice is a slice. The low end is one load in flight at a time. The high end is most of an engineer's week and several loads in parallel.
The full week, the rollover rule, and a calculator with the inputs exposed are on the capacity page.
Sprint Zero first. Its price credits against the first month.
A named engineer for part of the week. A small fleet compute budget, stated in dollars. One load in flight at a time. A Monday plan, daily commits, a Friday demo, a ship log.
Most of an engineer's week. A larger fleet budget, stated in dollars. Several loads in flight. Hotfix response the same business day. A quarterly architecture review in writing.
- rollover
- unused capacity, one month
- billing
- no hourly, no seats
Four things ship with every load.
Every paid load leaves with the same four things, whether it is a rescue, an agent, or a product. They are not add-ons and they are not a tier.
If one of them is missing at handover, the load is not done.
Evals
A test or eval harness that defines done in a way a machine can check, kept in the repo.
- lives in
- the repo
- checked by
- a machine, before merge
Observability
Logs, metrics, and an error monitor you can open without us.
- you get
- logs, metrics, an error monitor
- opened by
- you, without us
Runbook
A written runbook: how to deploy, how to roll back, what pages whom, what the bill is.
- covers
- deploy, roll back, who gets paged
- states
- what the bill is
30-day fix warranty
Anything we shipped that breaks within 30 days, we fix without a new quote.
- window
- 30 days from handover
- cost
- no new quote
What is not included.
A price without its exclusions is an anchor, not information. This is the list, per load, and it is the same list we will read back to you on the call.
- A read-only scan of the URL you paste, with findings ranked by how much they can hurt you.
- Nothing stored unless you ask us to send the result to your email.
- It does not log in, does not exercise your database policies, and does not read your code.
- It does not replace a person reading the repo. It tells you whether that person is worth calling.
- What is fine, what is dangerous, and what to do first, in writing after the call.
- If it fits an offer on this page, we say which one and what it would cost. If it does not, we say that.
- Not a sales call. No deck, no proposal unless you ask for one.
- Not a substitute for reading the code. That is a rescue estimate, below.
- One workflow, one integration, or one screen, deployed where you can use it.
- The fleet's task log, test results, and cost per task on a public build page.
- A follow-up call where we tell you what we would build next and what it would cost.
- Application-gated and capped at a few a month. Not every application is accepted.
- You agree to a public build log, with your name and product withheld if you prefer.
- Not a full product. A slice with an end, chosen so that seven days is honest.
- A one-page map of the workflows we can see from the call: what runs on a spreadsheet, a shared inbox, or a person's memory.
- A first-agent recommendation with the human approval gate drawn in.
- No dollar-per-month projection unless you give us the hours and the error cost; we do not invent them.
- Not a systems audit. One page, on purpose.
- Discovery, then a written spec with explicit non-goals.
- A working slice deployed, not a slide.
- An eval harness that defines done in a way a machine can check.
- An architecture and cost model, including token cost per task.
- A fixed-price quote for the load that follows.
- Not the whole build. The slice is real and deployed, but it is one slice.
- No fixed quote for the rest before the sprint ends, because the sprint is how we get one.
- A ranked list of what is wrong, in order of how much it can hurt you, before you pay for anything.
- Auth and data-layer fixes, migrations that run twice without damage, backups that restore.
- Tests on the paths that matter, so it stays fixed.
- CI, observability, cost controls, and a handover document.
- The option to move to a retainer, stated plainly in the handover.
- No forced stack migrations. If it runs on Supabase and Next, it stays on Supabase and Next unless the code itself is the problem.
- No quote before we have read the code. Read-only access is enough.
- No rescue without a scan first.
- No guaranteed launch dates. We give a plan with dates and we tell you the day it slips.
- An intake, support, scheduling, invoicing, or phone agent wired into the systems you already use: CRM, accounting, phone, email.
- Bounded autonomy: approval thresholds written down, and a person on anything irreversible.
- An audit trail of every action the agent took and why.
- Evals that run before every change, and a dashboard with a cost line.
- Tuning after it goes live, because the first month teaches everyone something.
- No agent for a thing that is just an automation. If a scheduled script does it, we build the script and say so.
- No autonomy on money, deletion, or outbound messages without a human gate.
- No per-seat pricing. You pay for the workflow, not for chairs.
- Product spec and a design system you can extend.
- A full-stack build with evals and CI, deployed on your infrastructure.
- Analytics, a security baseline, and the house standard: evals, observability, a runbook, a 30-day fix warranty.
- Then, if you want it, a standing slice of the machine to keep shipping.
- No fixed quote without a spec. Sprint Zero is how the spec gets written.
- No guaranteed launch date. A plan with dates, and the truth on the day one moves.
- Not everything in the roadmap. The first version your customers asked for.
- Permission-aware retrieval over the stores you have: SharePoint, Google Drive, email, ticketing.
- A citation on every answer and a log of the questions it could not answer.
- Evals that test retrieval separately from generation.
- No chatbot that answers from memory. If it cannot cite a document, it says it does not know.
- No bypass of your existing permissions.
- A named senior engineer who owns your roadmap and reads every diff.
- A monthly fleet compute budget, stated in dollars in the proposal.
- A Monday plan, daily commits you can see, a Friday demo, and a ship log.
- Rollover of unused capacity for one month.
- The house standard on everything shipped.
- No staff augmentation by the head. You buy shipped work, not a seat.
- No unlimited scope. The Monday plan is the scope for the week.
- Unused capacity rolls forward one month, not forever.
Against a loaded hire.
A senior software developer on US payroll runs about $20,000 a month once benefits are included, before recruiting, equipment, and management time (BLS OEWS May 2025 75th percentile wage of $171,980, times the 1.45 benefits load from BLS ECEC March 2026).
The top retainer sits under even the median developer's loaded cost. That is a comparison of a bill to a cost, not a claim about output: a retainer is a slice of the machine, not a person's whole week.
Bars scale to the 90th percentile loaded cost. The 75th percentile is the senior proxy; BLS does not split by seniority.
- formula
- loaded monthly = annual wage / 0.692 / 12
- wages
- Software developers (15-1252): median $135,980, 75th $171,980, 90th $214,670 a year (BLS OEWS May 2025).
- benefits
- 30.8% of total compensation, Private industry, professional and related occupations, Table 4, which multiplies wages by about 1.45 (BLS ECEC March 2026).
- excludes
- recruiting, equipment, and management time. The figure is a floor.
Tell us the task.
Two sentences from you. A human engineer reads every message and replies within one business day, Pacific time. You get a price and a plan, or the two questions we need answered first. No discovery funnel and no drip sequence.
Two sentences from you. A human engineer reads every message and replies within one business day, Pacific time.
Free. Paste a URL. No account.
Nothing on this page is a testimonial. What you can check right now: the scanner you can run on your own app, the fleet you can watch work, and the log of what shipped on this site with dates on it. See the scanner, the fleet and the log.