One drawing in.
Every downstream tool out.
Eight export formats, a REST API with Python and Node SDKs, webhooks, and an MCP server. One read, every register your team already runs on.
One read, every format.
The instruments read off the drawing land as the actual export file your downstream tool imports — the same four rows, rendered as a spreadsheet, a PLC vendor schema, or a process-industry exchange. Switch the format; the data does not move.
io-list.xlsx
| Tag | Class | Description | Signal |
|---|---|---|---|
| PT-101 | AI | Feed header pressure | 4–20 mA |
| FT-202 | AI | Feed flow | 4–20 mA |
| FCV-201 | AO | Flow control valve | 4–20 mA |
| XV-301 | DO | Inlet shutoff | 24 VDC |
Build with the API.
Bearer-key auth, idempotency keys, signed webhook payloads. The workspace and the API run on the same job queue.
POST /v1/extract
from tagsight import Tagsightts = Tagsight() # reads TAGSIGHT_KEYjob = ts.extract("unit-04.pdf") # upload the drawingjob.wait() # poll to completionio = job.results() # the classified I/O listio.to_excel("io-list.xlsx") # or .to_l5x() / .to_dexpi()
Reach Tagsight from your desktop AI client.
A Model Context Protocol server exposes seven Tagsight extraction tools to any MCP-compatible desktop client. Bluebeam Max ships MCP-native: drop a P&ID into the workspace, run the extraction, and pull the I/O list without leaving the drawing.