Building the pre-permit data layer for B2B real estate.

We're building the pre-permit data layer for B2B real estate.

cityminutes is building toward a 3,142-county planning-data layer — and extracts the four fields the rest of the category leaves on the table.

Our mission

For the 500,000+ Americans working in land acquisition, pre-construction, building product specification, and real estate development, the single most valuable piece of intelligence is the same: what a city is going to approve, reject, or condition six to twenty-four months from now.

That information already exists. It sits in planning commission minutes and city council agendas published by local governments under sunshine laws, open-meetings acts, and public-records statutes in every state. The problem is that each of the 3,142 US counties publishes on its own timeline, in its own format, on its own PDF portal, using its own vocabulary.

cityminutes exists to turn that distributed, unstructured public record into structured, searchable data — and to ship it as a product that a 25-person land acquisition team or a 300-person commercial GC's pre-construction desk can actually use. Pre-permit, transparent, and built for national coverage.

The problem we're solving

B2B real estate teams rely on stale, fragmented, post-permit data to make million-dollar decisions.

Stale. By the time a rezoning, variance, or subdivision hits a permit database, the planning commission has already voted. Conditions of approval are already recorded. Community objections are already registered. Staff recommendations have already shaped the outcome. The data underwriting a $50M site is, on average, 12–24 months older than the decisions that shaped it. Dodge's phone-reporter model, ConstructConnect's permit-stage feeds, and legacy lead-gen marketplaces all live downstream of what matters.

Fragmented. Roughly 90,000 local government entities in the US publish planning agendas, minutes, and hearing outcomes. Every one uses a different portal — Accela, OpenGov, CivicClerk, Granicus, Novus, IQM2, Legistar, or a PDF on a shared drive. Even the largest enterprise data teams can't cover more than a few hundred jurisdictions manually. The long tail is the product.

Post-permit. Every dollar of lead time lost at the pre-permit stage is a dollar of competitive advantage lost to whoever got there first. Developers who see a 14-condition list before the applicant's attorney revises it know whether the affordable set-aside kills the IRR — three months before the team with the Dodge subscription learns the project exists.

cityminutes makes the pre-permit layer structured and usable. That's the whole product.

Our approach

Four-part stack. None of the parts are novel alone. The combination is.

1. Automated ingestion. We are building adapters for planning and council agendas across major platforms (Accela, OpenGov, CivicClerk, Granicus, Novus, IQM2, Legistar, and custom municipal sites). Built for the long tail — the counties that publish as PDFs-on-a-filesystem, not just the places with a public API.
2. AI-assisted extraction. Large language models (with human-in-the-loop review on the 4 wedge fields) extract structured data: conditions of approval, community objections, hearing outcomes, staff recommendations, rezoning, variances, CUPs, subdivision plats, site plans, development agreements, plus full contextual metadata.
3. Human QA loop. Every wedge-field extraction is spot-checked against the source. We publish accuracy metrics on /trust and link every structured field back to the primary source.
4. Public-indexed content. County, city, operational-query, and glossary pages are server-rendered, canonical-tagged, and schema-decorated. LLMs can crawl us. Google can index us. Journalists can cite us. This is the structural advantage boardwalkai ceded by auth-gating their county pages.

Plus transparent pricing — published tiers on /pricing, no 60-day procurement gauntlet.

Our values

Transparency. No PBN backlinks. We publish our coverage list, accuracy metrics, vendor list, and changelog on public URLs. When we make a mistake, we correct it publicly.

Data integrity. Every structured field links back to the primary source. Every record has an audit trail. If a customer disputes a field, we cache the source document and roll back.

Accessibility. Public-indexed content means researchers, journalists, and civic-tech users can read our county pages without an account. Our Data With Purpose commitment — free nonprofit and academic tier modeled on the regrid partnership program — makes the pre-permit layer available to Trust for Public Land, UC Berkeley SGMA, National Zoning Atlas, and Conservation Fund–style users.

Builder-focused. B2B only. No retail. No house flippers. Every product decision optimizes for the four personas: land acquisition, pre-construction BD, building product spec reps, and real estate developers and investors.

Nationwide target. 3,142 US counties and county-equivalents are the product target, with current coverage and roadmap details on /coverage.

The team

cityminutes is founder-led while the product, data, and go-to-market team is built out. Named team bios will be published after roles and public profiles are finalized.

Company milestones

• 2026 Q1 — cityminutes founded. Pre-permit data layer thesis locked.
• 2026 Q2 — 20-county launch target. First 4 AEO wedge pages published. Regrid partnership conversation opened.
• 2026 Q3 — 100-county milestone. First enterprise contract signed. Data With Purpose nonprofit tier announced.
• 2026 Q4 — 500-county milestone. Procore Marketplace listing evaluated. First quarterly data report published.
• 2027 Q1-Q2 — Nationwide coverage target. Full 4-field wedge available across target coverage. SOC 2 Type II target.

Timeline is a commitment, not a prediction. Updates ship on /company#milestones and the public changelog.

CTA

• Primary: See the product →
• Secondary: See coverage →