How we make recommendations.

What happens when you generate a plan

You pick a situation, a project type, and up to three concerns. We match those answers against a published rule set — each rule contributes questions to ask, documents to request, organizations to contact, and red flags worth knowing.

The result is a single composed plan: deduplicated, ordered into a 24-hour / this-week / ongoing timeline, and stamped with a confidence note. Nothing is sent to a server. Nothing is stored anywhere outside your browser.

What we deliberately don't do

• We don't predict outcomes. Local decisions depend on dozens of factors we can't see.
• We don't give legal advice. Easements, appeals, and zoning challenges deserve a qualified attorney.
• We don't fingerprint you. There's no account, no tracker, no analytics service.
• We don't surface "match scores" or fake confidence. If a plan is general, we say so.

Anonymous improvement metrics

The feedback block at the bottom of generated plans is opt-in and lives on your device until you choose to send it. No automatic telemetry. We use only what you intentionally share, and only to make the guide clearer.

How often this is reviewed

The rule set has a last reviewed date in the source. Resources and glossary entries carry the same date. When a rule changes, we say what changed and why.