PLSS, built for AI
Two ways AI works with Township America — chat with our in-app assistant, or connect your own agent via the MCP server. Same accurate, current PLSS data either way.
AI assistant works on all plans. MCP server bundled with Pro+ ($99/mo) and Business.
Ask the AI Assistant
- Convert PLSS in chat
- Batch + route planning in plain English
- Federal Land queries on demand
- Plain English place-name search
Daily message allowance
Connect Your Own Agent
- 6 tools: search, coordinates, GeoJSON, validate, batch, land report
- Works with Claude Desktop, Cursor, Continue, Cline
- X-API-Key auth, 1K calls/mo bundled
- npm + PyPI distribution
MCP access by plan
Capabilities
What each surface can do
In practice
Three scenarios where PLSS AI earns its keep
Landman, mineral rights research
Ask Claude: “Find all active O&G leases in T2N R68W, 6th P.M., Weld County, then export polygons.” Claude uses the Township America MCP server, returns BLM lease data and a Shapefile — no manual steps.
GIS analyst, batch conversion
Open Cursor with the MCP server connected. Paste 500 PLSS descriptions. The agent batch-converts via the bundled batch tool and writes results straight back into your codebase or notebook.
Title agent, quick lookups
Ask in the Township America web app: "What's the section for 1234 N. Main St, Caldwell, ID?" Get back the legal location, county, principal meridian, and a map preview — no form to fill out.
Data & Accuracy
Both surfaces run on the same data
Current BLM CadNSDI V2, refreshed quarterly. 30 PLSS states, 37 principal meridians. The Assistant and MCP server call identical underlying endpoints — no separate data path, no sync lag.
- BLM CadNSDI V2 — refreshed quarterly
- 30 PLSS states, 37 principal meridians
- Claude (Anthropic) powers the in-app Assistant
- MCP server open-source at github.com/townshipamerica/typescript-mcp
- Sub-200 ms REST API response times powering both surfaces
Plans
AI features by plan
Frequently asked questions
What's the difference between the AI Assistant and the MCP server?
The Assistant is a chat surface inside the Township America web app — you sign in, ask questions, get answers. The MCP server lets your own AI tools (Claude Desktop, Cursor, Continue, Cline) call Township America's PLSS functions directly. Same data, different consumption pattern.
Do I need to be a developer to use MCP?
Mostly yes — MCP requires editing a JSON config file in your AI client (Claude Desktop, Cursor, etc.) and supplying an API key. The Assistant requires no setup beyond a Township America login.
Does the AI replace the REST API?
No. The REST API is the raw HTTP interface; the AI surfaces (Assistant + MCP) sit on top of it. Power developers may want both — use the REST API for high-volume batch jobs, MCP for interactive agent workflows.
Which models power the in-app Assistant?
The Assistant uses frontier models from Anthropic (Claude). Model selection may change as new releases ship; the user experience stays the same.
Are my queries stored or used for training?
Queries are logged for product analytics (usage tracking, billing). They are not used to train any AI model. Township America has no training pipeline.
Can I use my own API key with my own AI model?
For MCP: yes, you can call our REST API directly from any agent framework via the X-API-Key header. For the in-app Assistant: no, the model and provider are fixed.
Get started
Two AI surfaces. One PLSS dataset.
Try the in-app Assistant today, or wire up your agent via MCP. Both ship on Pro+.