opentakeoff-mcpCAUTION
Open-source (AGPL-3.0) PDF takeoff for construction & flooring — the first engine an AI agent drives natively over MCP, not bolted on. Rooms measured on the wall faces, symbol sweeps, materials + quantities with full provenance, built for preconstruction. Runs entirely in your browser.
Overview
From the repository's own README, as read at the audited commit. Badges and raw HTML are left out.
The measurement engine for building plans—built so an AI agent can drive it, and so an estimator wants to.
A takeoff is the act of measuring quantities off a construction drawing. OpenTakeoff does it two ways over one engine: 53 MCP tools for an agent, and a browser canvas for a person. Agents and people share the takeoff document and quantity calculations. Each sheet carries its calibration; measurements carry geometry, method and authorship. Recalibration updates quantities together, incompatible imports report scale conflicts, and agent measurements carry an explicit review status. See the Phase 1 test guide.
[](LICENSE) [](https://opentakeoff.kentucky-ai.com) [](https://registry.modelcontextprotocol.io) [](https://www.npmjs.com/package/opentakeoff-mcp) [](https://union.kentucky-ai.com) [](https://openarena.to/en/projects/cmsgykvsq0000mkuv7byhlgnl) [](https://github.com/sponsors/Kentucky-ai)
For agents · **Try the canvas** · The engine's contract · For the person at the canvas · The data layer · Research · Fork it · Contribute
**Th
a98be0c9df86OBSERVED · 2026-10-12Connect
Built from this server's own package name, version and transport as found in its source — not copied from anyone's documentation, so it cannot drift against a page we do not control. Replace the environment placeholders with a token scoped to the least it needs.
claude mcp add opentakeoff-mcp -- npx -y [email protected]
Exposed tools (63)
45 read · 15 write · 3 destructive. Blast radius: 3 tools can delete or overwrite — an agent that can be talked into calling a tool can be talked into calling this one.
| Tool | Risk | Description |
|---|---|---|
annotate | read | |
apply_rules | write | |
count_marks | read | |
create_condition | write | Create a new takeoff condition for a finish tag (e.g. CPT-1) when no existing condition matches. Returns its condition_id. |
create_rfi | write | |
cut_out | read | |
delete_rfi | destructive | |
delete_shape | destructive | |
delete_verdict | destructive | |
derive_base | read | |
derive_transitions | read | |
detect_rooms | read | |
duplicate_condition | read | |
edit_annotation | write | |
edit_condition | write | |
edit_materials | write | |
edit_shape | write | |
export_dxf | read | |
export_marked_pdf | read | |
export_report | read | |
export_takeoff | read | |
find_schedule | write | |
find_text | read | |
get_conditions | read | List the takeoff conditions (finish tags) that exist in this workspace, with their ids. Proposals must reference an existing condition_id. |
get_sheet_vectors | read | |
import_takeoff | write | |
link_annotation | read | |
list_annotations | read | |
list_rfis | read | |
list_shapes | read | |
list_sheets | read | List the sheets open on the canvas: key, title, pixel dimensions, and scale status. Tools only work on open sheets. A sheet without a scale set cannot be measured — say so and stop rather than guessing. |
load_plan | read | |
look | read | look at something |
mark_verdict | read | |
measure_line | read | |
measure_polygon | read | |
measure_surface | read | |
one_click | read | |
open_tool_stage | read | |
place_count | read | |
probe | read | probe something |
propose_condition_edit | write | |
propose_takeoff | read | |
read_schedule | write | Read a finish/material schedule table inside a region of a sheet into structured rows — Import from schedule |
read_sheet_text | read | |
resolve_rfi | read | |
resolve_tag | read | |
revise_proposal | read | |
scope_duplicates | read | |
scope_merge | write | |
set_scale | write | |
sheet_context | read | |
sheet_graph | read | |
sheet_info | read | |
split_condition | read | |
sweep_schedule_row | write | |
symbol_sweep | read | |
takeoff_summary | read | |
undo_last | read | |
view_region | read | Render a region of the sheet as an image and look at it. Use this for raster sheets, hatched/ambiguous areas, or to visually confirm what a room contains before proposing. |
view_sheet | read | |
withdraw_condition_edit | write | |
withdraw_proposal | read |
Trust audit
CAUTIONgrade C · trust 74/100 Install with care. The audit found things worth knowing before you trust its output.
| Layer | What it checks | Result |
|---|---|---|
| L0 | Provenance & inventory | PASS |
| L1 | Static analysis of the code | WARN |
| L2 | Instruction surface (what it tells the agent) | PASS |
| L3 | Class-specific surface | WARN |
| L4 | Behavioural (sandbox) | SKIPPED |
What the source does
- Filesystem
- declared (7 observation(s))
- Network
- declared (8 observation(s))
- Shell
- declared (3 observation(s))
- Dependencies
- not all pinned
- Secrets in source
- none-found
Findings (25)
mod = importlib.import_module(mod_name)
base = f"http://127.0.0.1:{httpd.server_address[1]}"const ver = await (await fetch(`http://127.0.0.1:${cdpPort}/json/version`)).json();const BASE = 'http://127.0.0.1:5261/';
delete_rfi, delete_shape, delete_verdict
s1-noisy-p01.wav
s1-noisy-p02.wav
s1-noisy-p07.wav
s1-noisy-p10.wav
s1-noisy-p12.wav
mq-c-epox-19-c1-c2.json
return hashlib.sha1(json.dumps(key, sort_keys=True, default=str).encode()).hexdigest()
phash = hashlib.sha1(blob.encode()).hexdigest()[:12]
import { Session } from "../../../../mcp/src/session.ts";const DOCS = ["../../README.md", "../../FEATURES.md", "../../AGENT_BRIEF.md",
"../../README.zh-Hans.md", "../../README.ja.md", "../../README.ko.md",
"../../docs/USER_GUIDE.md", "../README.md", "../../docs/MCP.md", "../../docs/AGENT_GUIDE.md",
await import("../../scripts/check-version-consistency.mjs");Open http://127.0.0.1:5187 in a disposable browser profile for testing. Existing project data in another profile is independent.
Observed: web check passed (1,846 tests: 1,843 passed, 3 skipped, 0 failed), including typecheck, lint, benchmark, and production build. MCP inventory and wiki checks passed. Protocol: 61 passed, 0 fa
const setFileInPage = async (selector, file, mime) => { const { readFileSync } = await import("node:fs"); const b64 = readFileSync(file).toString("base64"); const name = file.split("/").pop(); await e@modelcontextprotocol/sdk, @turf/boolean-contains, @turf/boolean-overlap, @turf/difference, @turf/helpers, jsts, pdf-lib, pdfjs-dist
tsx
pytest, httpx
fastapi, uvicorn, pydantic
Gates applied: no_behavioural_pass.
a98be0c9df86full audit observations/trust-audit/mcp-server/kentucky-ai__opentakeoff.json · Report an issue / request a re-scanAudit history
Every audit this server has had. A grade with a past is a grade somebody is still checking.
| Date | Source | Verdict | Grade | Score | Change |
|---|---|---|---|---|---|
| 2026-10-12 | a98be0c9df86 | CAUTION | C | 74 | first audit |
Questions
What is the opentakeoff-mcp MCP server?
Open-source (AGPL-3.0) PDF takeoff for construction & flooring — the first engine an AI agent drives natively over MCP, not bolted on. Rooms measured on the wall faces, symbol sweeps, materials + quantities with full provenance, built for preconstruction. Runs entirely in your browser.
What tools does opentakeoff-mcp expose?
63 in total: 45 read-only, 15 that write, and 3 that can delete or overwrite (delete_rfi, delete_shape, delete_verdict). Every one is listed on this page with its risk.
Is opentakeoff-mcp safe to connect to an agent?
With care. The audit graded it C (74/100) and found 25 things worth knowing before you trust this server, listed below with the exact line each was found on. Separately from the audit: 3 of its tools can destroy data, so scope the token you give it to what you actually need.
What credentials does opentakeoff-mcp need?
It reads OT_SANDBOX_API_KEY and OT_SECOND_READ_KEY from the environment. Give it a token scoped to the least it needs — an agent that can be talked into calling a tool can be talked into calling it with your credentials.
How does opentakeoff-mcp run?
It speaks stdio, so it runs as a local process your client starts. It is published on npm as opentakeoff-web at 0.1.0.
How current is this page?
The grade is for one exact copy of the source (a98be0c9df86), read on 2026-10-12. The repository is watched and re-audited when it changes.