Energy land, checked. Solar, wind, battery, and data-center projects, the permits they need, and the grid they plug into. Start here →
Parcel & Grid

Home › For developers › Diligence Reader

Diligence Reader

Reads development documents with your own Anthropic API key · Nothing is sent to Parcel & Grid · Built October 18, 2026

Drop in the documents a project has so far: the critical issues analysis, the ALTA survey, the title commitment, the option or lease, the IPaC letter, the interconnection study. For each one it pulls out the constraints with their numbers, legal access, permits and interconnection items mentioned, dates, costs, open items, and the study or filing the document points to next. Across the set it shows where the documents disagree, which standard diligence documents are still missing, and one dated timeline.

How it works, and what leaves your computer. The page is static. When you press Read, your browser sends each file straight to api.anthropic.com with your key, under the terms of your own API account; the result comes back to your browser and is shown here. Parcel & Grid has no server in the path and keeps nothing. Your key stays in this page's memory unless you tick "remember in this browser", which stores it on your own device. Do not upload documents you are not permitted to send to a third-party service; most consultant reports and title commitments carry confidentiality terms, and your employer's documents are your employer's call.
Keys are made at console.anthropic.com. Cost depends on the model and the page count; your console shows the exact usage after each read.

Drop PDFs, images or text files here, or choose files. Up to 30 MB per file (this page's limit) and 100 pages per PDF (the API's limit); export Word files to PDF first.

    One request per file; a file is read once unless you press Read again.

    What it extracts

    Limits

    An AI model reads the document; it can misread a table, skip a page in a long scan, or state a finding the document only implies. Every item should be checked against the source before it goes into a memo, and the extraction is not a legal review. People's names are left out by instruction; company and agency names are kept. Scanned PDFs without a text layer are read as images, which works for clean scans and poorly for faint ones. The reader does not run the public-data screen; for that, the land report and its Desktop CIA export cover the point you pick.

    Not a substitute for the advice of an attorney. See the terms and the privacy policy (section 1 covers this page).

    Browse resources