Appointment Feasibility API

The Appointment Feasibility API answers one question booking software can't get from a maps provider: which appointment slots can a technician actually reach? Send the day's existing jobs, a time window, and the new job's location — it returns only the start times that work, accounting for real road-network drive times, travel buffers, and an optional home base. Full-US coverage on monthly-refreshed OpenStreetMap data.

How does it work?

One POST with the schedule as it stands; the response is the bookable truth — no slot is returned unless the tech can finish the previous job, drive over, do this job, and reach the next one:

POST /stl-api/v1/drivetime/feasibility
{"job": {"lat": 38.88, "lon": -94.81},
 "window_start": "2026-07-14T08:00:00-05:00",
 "window_end":   "2026-07-14T18:00:00-05:00",
 "duration_minutes": 90,
 "appointments": [{"start": "…", "end": "…", "lat": 39.0, "lon": -94.6}],
 "home_base": {"lat": 38.9, "lon": -94.7}}

{"feasible_slots": [{"start": "…", "end": "…",
  "drive_in_seconds": 1140, "drive_out_seconds": 1260, …}], "count": 6}

This is the engine behind "don't book a 2pm in Olathe when the tech is in Lee's Summit until 1:45" — as an API, without building the gap-and-buffer math yourself.

Why not just use Google Maps or Mapbox?

Use them — if raw travel times are all you need; their free tiers are generous. They sell elements; you still own the hard part: turning travel times into bookable slots around a live schedule. This API sells the finished decision. Two more differences that matter in production: a monthly hard spending cap (Google and Mapbox bill uncapped — alerts, not stops), and the same key doubles as an MCP server so AI agents get these as native tools.

Point-to-point drive time (included)

The supporting utility, when you do want a raw answer:

curl -s https://api.composedchaos.net/stl-api/v1/drivetime/route \
  -H "X-API-Key: $KEY" -H 'Content-Type: application/json' \
  -d '{"origin":{"lat":39.0997,"lon":-94.5786},
       "destination":{"lat":38.9884,"lon":-94.6708}}'

{"seconds": 1500, "meters": 18227, "source": "valhalla"}

What does it cost?

$49/mo Developer plan — 10,000 drive-time queries (feasibility, routing, service-area) and 200 lead-email parses included monthly, then $0.01/query and $0.02/parse. Monthly hard cap at 3× the included allowance (adjustable on request): a runaway agent loop can never produce a surprise bill. No setup fee, cancel anytime.

Get a key

Get an API key now — shown instantly, activates when checkout completes.

FAQ

Can AI agents use it? Yes — the same key is the bearer token for an MCP server (official registry: net.composedchaos/speed-to-lead); Claude and any MCP-capable agent get feasibility, drive time, service-area verdicts, lead parsing, and a lead store as native tools with one config block.

Coverage? Full United States, monthly-refreshed OpenStreetMap.

Rate limits? 60 drive-time queries/min, published in the API reference, with 429 + Retry-After past them. Your live usage and caps are always visible at GET /v1/metrics.

Full reference: api.composedchaos.net/docs