Hand over the stops, get back the order that visits them in the least time, plus the turn-by-turn route through them. Start and end stay where you put them; everything in between is solved.
10,000 units free every month. No credit card, no sales call.
Seven drops in the order they were typed in.
POST /optimized-route/v1 · 9 locations · costing: bicycleinput order · 47.4 km · 2 h 56
Nobody ships an endpoint. They ship a run sheet a rider can follow, a van that is back before the school run, a field-service app that stops asking engineers to plan their own day. This is what a solved order looks like from the outside.
As entered47.4 km · 2 h 56
Optimised31.7 km · 2 hSeven drops around Amsterdam, entered in the order they came in. The solved order visits the same seven and comes home almost an hour earlier. This is one real round, measured, not drawn.
The same road graph, routedThe response comes back in the solved order with a leg per hop, so the list your rider follows is the response itself: next stop, minutes to get there, repeat.
Then navigate each legA round gaining stops, solved once each morning
1 unit per stop, up to 20
the matrix alone would be 4
An optimisation costs one unit per stop, up to twenty stops a request. Solving it yourself would start with the full travel-time matrix, which is stops squared.
See what a month costsconst { trip } = await rijwind.optimizedRoute({ locations: [depot, ...drops, depot], costing: 'bicycle', }); // the solved order, as your own stop ids const order = trip.locations.map((l) => drops[l.originalIndex - 1]);
The response has the shape of a plain route, with your stops echoed back in the solved order and an originalIndex on each. Swap the endpoint and the code that draws the line does not change.
Read the quickstartThe response is a plain route: your stops echoed back in the solved order, each carrying the index it had in your request, then a summary and a leg per hop. The code you already have for drawing a route draws this one.
Everything a developer scans a product page for, in one place, without the adjectives.
Numbers as of September 2026. The pricing page has the full ladder.
You give it a start, an end and the stops in between; it returns the order that visits all of them in the least travel time, together with the full route through them. The first and last locations stay where you put them, so a round that starts and ends at the depot stays a round.
Plain routing visits the stops in the order you gave them. Optimisation chooses the order. On the round on this page that is the difference between 47 and 32 kilometres for the same seven drops, because the entered order hopped back and forth across the city.
One unit per location in the request, from the same monthly pool as maps, geocoding and routing. Nine locations is nine units. Solving the order yourself would start with a travel-time matrix, which costs the square of that.
Between two and twenty locations, start and end included. Larger rounds are usually split by area or by vehicle first, then each part is optimised on its own.
No. It solves the visiting order for one vehicle with no constraints beyond start and end. If you need delivery windows, load limits or a fleet, this is the building block under that logic, not the whole scheduler.
Car, bicycle and pedestrian, each with its own rules, so a bike round is solved on the roads a bike can use.
No. Durations come from the road graph (speed limits, road class, surface and turn costs), not from live traffic. Rijwind Traffic is a separate live layer for your map and does not feed into the solved order today.
The same shape as a plain route: your locations in the solved order, each with the index it had in your request, a summary with total distance and time, and a leg per hop with its geometry and manoeuvres.
Yes. The free plan includes 10,000 units a month with no card and no sales call, which is a nine-stop round a thousand times over. It is a hard cap, so a runaway loop cannot produce a bill.
In the European Union, end to end. No part of the request path touches a US-controlled provider, which is what makes the GDPR and data-residency conversation short. We publish a DPA and a full sub-processor list.
Turn-by-turn for each leg of the solved round.
Every source to every target, when you solve it yourself.
A noisy GPS trace, snapped to the roads it was on.
Turn the addresses on the run sheet into coordinates.
One unit pool, one table, a cap you set.
What actually differs, feature by feature.
Ten thousand units a month, free. The same key covers maps, search and routing when you get to them.