mcpField
G
ground-motion

ground-motion

Remote
sibfly.comScience & Education0Updated 17w ago

About ground-motion

ground-motion is a hosted API for measured vertical ground movement at US addresses, based on NASA OPERA Sentinel-1 InSAR data. It serves screening estimates with uncertainty for real-estate, insurance, infrastructure and geotechnical use.

Key features

  1. 01Measured vertical movement in mm/year
  2. 02Address-level 30 m grid results
  3. 03Uncertainty, confidence and provenance
  4. 04Coverage and cost preflight
  5. 05Batch requests and webhooks

Use cases

  1. 01Screen a home for possible subsidence
  2. 02Review properties for underwriting
  3. 03Monitor infrastructure portfolios
  4. 04Pre-screen sites for geotechnical work

MCP support

The hosted Streamable HTTP MCP server at https://sibfly.com/mcp exposes nine tools for ground-motion checks, coverage, history, account and payment operations. Clients can also register through the server.

Configuration

Remote servers need no local install. Pick your client, then copy the config or install in one click.

Endpointhttps://sibfly.com/mcp
Copy install command
Claude CodeTerminal
Cursor.cursor/mcp.json
VS Code.vscode/mcp.json
Claude Desktopclaude_desktop_config.json
Windsurf~/.codeium/windsurf/mcp_config.json

Run commands in a terminal; paste JSON into the client’s MCP config and restart. Servers that need sign-in will prompt on first connect.

Tools

9
ParameterTypeRequiredDescription
latnumberNo
lonnumberNo
sincestringNoYYYY-MM-DD; if no new data since this date, return free
addressstringNoUS street address
compactbooleanNolean payload (fewer tokens)
dry_runbooleanNoFREE preview: coverage + would_cost + data age, no rate, no charge
explainbooleanNoadd ground_brief: ~600 tokens of plain-English narration + FAQ (engine GL52 free tier / CL46 paying). Adds ~10s latency. Only set when a human will read the output; use ask_inspector for follow-up Q&A.
max_age_daysintegerNoif data older than this, return free instead of billing
min_confidencenumberNo0-1; below this, return free instead of billing
require_significantbooleanNoif the rate would be no_clear_trend (within its error bar), return FREE instead of billing — never pay for statistical noise

Details

From the project README

SibFly — measured ground subsidence API (mm/year) for any US address

Is the ground under an address sinking? SibFly returns measured vertical ground motion — subsidence and uplift, in millimeters per year and inches per year — for any US address, from NASA OPERA Sentinel-1 InSAR satellite radar. Measured, not modeled.

Quickstart

# 1) an AI agent can register itself — no human, no captcha
curl -X POST https://sibfly.com/api/v1/autonomous/register \
  -H "Content-Type: application/json" -d '{"email":"YOUR_EMAIL"}'

# 2) measure the ground under an address
curl "https://sibfly.com/api/v1/motion?address=1100+Congress+Ave+Austin+TX" \
  -H "Authorization: Bearer sf_live_..."

Returns velocity_vertical_mm_yr / _in_yr, an uncertainty range, assessment_code, confidence, provenance, and cost. Flat $0.40 per covered report; misses are free; free coverage/cost preflight via ?dry_run=1.

What it is

  • Data: NASA OPERA DISP-S1 (Copernicus Sentinel-1 C-band InSAR), 30 m grid, 2016–present.
  • Vertical mm/year of ground movement (not elevation), measured, per exact 30 m cell, with uncertainty.
  • Coverage: US (expanding). Not real-time (source refreshes ~every 12 days); a screening estimate, not survey-grade.

For AI agents

Self-registration, free coverage/cost preflight, idempotency keys, per-key spend caps, scoped browser-safe keys, batch, webhooks, x402 (USDC on Base), test-mode keys, and a hosted MCP server with 9 tools: check_ground_motion, check_portfolio, get_coverage, get_motion_history, get_account, get_usage, ask_inspector, register_agent, buy_credits. The full, authoritative contract lives at https://sibfly.com/llms.txt and https://sibfly.com/openapi.json — always current.

Connect from an MCP client (Cline, Claude, Cursor, any Streamable-HTTP client)

SibFly is a hosted (remote) MCP server — nothing to install or run locally.

  1. Get an API key (one call, free credits included, no captcha):
    curl -X POST https://sibfly.com/api/v1/autonomous/register \
      -H "Content-Type: application/json" -d '{"email":"[email protected]"}'
    
    Save the returned api_key (shown once). Or skip this step: connect without a key and call the register_agent tool from inside MCP.
  2. Add the server to your MCP client config (Cline: MCP Servers → Remote Servers, or cline_mcp_settings.json):
    {
      "mcpServers": {
        "sibfly": {
          "type": "streamableHttp",
          "url": "https://sibfly.com/mcp",
          "headers": { "Authorization": "Bearer sf_live_YOUR_KEY" }
        }
      }
    }
    
    OAuth 2.1 (Dynamic Client Registration + PKCE) is also supported for clients that prefer a "Connect" flow — discovery at /.well-known/oauth-authorization-server.
  3. Try it: call get_coverage (free) on any US address, then check_ground_motion ($0.40, misses free). dry_run: true previews coverage, cost, and expected answer quality without billing. Out of credits? The buy_credits tool returns a payment link (card or crypto).

Sandbox: register with {"mode":"test"} for an sf_test_ key — $100 fake credits, synthetic data, rehearse the whole flow for $0.

Use cases

Real-estate due diligence · "is my house sinking?" for home buyers · insurance & mortgage underwriting screening · infrastructure and portfolio subsidence monitoring · geotechnical pre-screening.


A product of SIB Scientific. Screening estimate from satellite radar (NASA OPERA / Sentinel-1 InSAR) — not an engineering, survey, or hazard determination.

Comments

Markdown supported

New to MCP?

From what the protocol is to installing your first server, these answers cover most first questions.

An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.