Company profile · 3 min read
Cranston AI: tax prep software for CPA firms
Cranston combines AI tax prep, bookkeeping and practice management for CPA firms while keeping reviewers in control.
Cofounder of Lazyweb (a16z Speedrun). Previously grew subscriptions and built native ads at Duolingo. Writes First1000 (90,000+ subscribers).
Published · Updated
Cranston is tax prep and bookkeeping software for CPA firms. It fits firms that want to increase capacity without replacing QuickBooks, Drake, CCH Axcess, Lacerte or UltraTax—and still want a reviewer to make the final call.
What it does
The workflow is unusually concrete. Cranston’s homepage says it reads source documents, enters returns in the tax software, attaches workpapers, flags missing or wrong-year documents, drafts client follow-ups, classifies transactions, reconciles accounts and queues exceptions. The practice-management layer keeps each client’s status in one place. Nothing posts, files or reaches a client until a reviewer approves it.
Why I’d look closer
That reviewer boundary is the product’s strongest decision point. This is not a promise that the AI replaces the CPA. It is a promise that the CPA spends less time typing, chasing and reconciling. The YC profile says Max Minsker previously owned a tax practice that filed 10,000+ returns; it describes Sean O’Bannon as a Stanford CS/AI founder who previously built ReMatter and worked at Databricks. Those are company-profile facts, not an independent endorsement.
The pricing is refreshingly legible. Cranston’s pricing page lists $30 per client book per month, $40 per prepared return and practice management at $0. A growing firm can estimate the bill from books and returns instead of negotiating a seat bundle. The company also says on its security page that its SOC 2 Type II examination is in progress. That status matters if the buyer handles sensitive taxpayer data.
What I’d ask
The public search surfaced Cranston’s company LinkedIn page, not an independent review. That is useful for confirming the company’s public positioning; it is not evidence that the workflow works for every firm. I would still ask about migration effort, exception rates, supported return types, audit trails and the exact boundary between Cranston’s preparation and a CPA’s responsibility.
My editorial take
Shortlist Cranston for a CPA firm with repeatable volume and a real review process. It is less obviously suited to a solo preparer who wants a simple tax chatbot. The interesting bet is not “AI does taxes.” It is that a firm can keep its existing systems while moving the repetitive work—and the evidence around that work—into a reviewable queue.
Quick facts
| Field | Sourced detail |
|---|---|
| Accelerator | Y Combinator, Fall 2025 |
| Industry | B2B; Finance and Accounting |
| Location | San Francisco |
| Public website | cranston.ai |
| Pricing | $30 per client book/month; $40 per prepared return; practice management included |
| Constraint | Company says its SOC 2 Type II examination is in progress |
Sources checked
| Volatile field | Source | Checked |
|---|---|---|
| Membership/status/founders | YC profile | 2026-09-19 |
| Pricing | Cranston pricing | 2026-09-19 |
| Security status | Cranston security | 2026-09-19 |
| Public company search | Cranston LinkedIn | 2026-09-19 |
Cohort context
Cranston AI is listed in Fall 2025. In our 2026-09-18 directory snapshot, 90 of 146 listed companies in that cohort have YC’s primary industry label B2B (61.6%). This is a current-directory comparison, not an original intake count or a performance ranking. Nine-cohort dataset.
Public website snapshot
Observed 2026-09-19T16:14:58.989Z in raw homepage HTML. This records visible metadata and advertised links, not agent execution or product quality.
| Signal | Homepage observation |
|---|---|
| Product description metadata | Observed |
| Canonical link | Observed |
| H1 or H2 heading | Observed |
| Typed structured data | Observed |
| Docs/developer link | Not observed in this response |
| Pricing link | Observed |
| llms.txt link | Not observed in this response |
| Markdown alternate | Not observed in this response |
Public observations · Collection method. Missing links here do not establish that a capability or file is absent elsewhere.
About the author
Cofounder of Lazyweb (a16z Speedrun). Previously grew subscriptions and built native ads at Duolingo. Writes First1000 (90,000+ subscribers).
I cofound Lazyweb and publish Mudpie. This is an owner-written publication, not an independent testing organization. Research notes distinguish observations, sourced reporting and editorial judgment.
First1000 ↗ · X ↗