Where things stand

The idea

Forestry operations already generate detailed machine and contract data — from harvesters, forwarders, and delivery records. Forestry AI's goal is to let the people who actually need that data (sellers, operators, machine owners) just ask for it in plain language, instead of learning a reporting tool. The three examples on the home page are drawn from real StanForD machine-data concepts, not invented ones.

This isn't being built to launch as a company. The aim is to bring it — and the engineering behind it — to a forestry business as part of the team, not sell it to one.

Public vs. private

The engineering behind Forestry AI is split deliberately:

• forestry-sdk-for-net Public .NET SDK: streaming StanForD document parsing and the conversational pipeline that turns a question into an answer.
• forestry-sdk-for-python Public Python SDK: a terminal client used during development to exercise the core system end-to-end.
• Chat (private) The core natural-language-to-answer logic. Kept private while the approach is still being proven out.

Timeline

• In progress
  Core pipeline (private): turning a question into a structured intention, then an answer.
• In progress
  Public SDKs: streaming document parsing, schema reflection, and the client-facing pipeline — built in the open so the engineering is visible on its own.
• Planned
  A real StanForD schema project (harvesting, forwarding, and quality documents), replacing the placeholder fixtures used in testing today.
• Planned
  A working demo connecting a real reader to real data, beyond the mocked examples shown here.