Geospatial search,built for speed.

Q: How fast is sub-millisecond, really?

Median radius queries against a 10M-point dataset return in 0.3–0.7ms on commodity hardware. See the benchmark page for full numbers, methodology, and source code.

Q: Where does my location data live?

On your worker, on your hardware. The Orter control plane only stores collection metadata and routing info — your raw geospatial data never leaves your network.

Q: What interfaces does the worker expose?

REST (HTTP+JSON), gRPC (Protobuf), and a local UNIX socket. Pick whichever fits your stack. SDKs for TypeScript, Python, and Rust are also available.

Q: Can I run this on-prem or air-gapped?

Yes. The worker is a single Docker image. It can run fully offline once provisioned. The control plane is optional for advanced features.

Q: Is Orter open source?

The Rust-powered worker core is open source. The control plane and dashboard are proprietary, free up to the generous free-tier limits.

Q: What does pricing look like?

Free forever for small workloads. Paid tiers scale by queries-per-second and number of worker nodes. No per-record fees, no surprise bills.

Find the nearest location across millions of points in sub-millisecond time. Deploy a worker anywhere, query via REST, gRPC, or UNIX socket.

No credit card

Free forever

scroll

Built for production geospatial workloads

< 1msavg query latency

REST · gRPC · UNIXquery interfaces

On-premisesself-hosted worker

Open SourceRust-powered core

Any hardwarePi → production

Why Orter

Stop fighting your geospatial stack.

Most geospatial setups are slow, bloated, or locked into a single cloud. Orter fixes all three.

The old way

Postgres + PostGIS tuning nightmares

Managed services that lock you in

Hundreds of ms per radius query at scale

Sending location data to a third party

Complex setup, weeks to production

The Orter way

One Docker command to deploy a worker

Run anywhere — your VPC, on-prem, or a Pi

Sub-millisecond queries by default

Your data never leaves your network

Production-ready in under 3 minutes

How it works

Up and running in minutes.

No infrastructure expertise required. Three steps from zero to sub-millisecond geospatial queries.

Upload your data

Add location datasets to Orter collections via the web dashboard or API. Any format, no schema changes needed.

Deploy a worker

Run a single Docker container anywhere — your server, a cloud VM, or even a Raspberry Pi. One command, zero config.

Query at speed

Call via REST, gRPC, or UNIX socket. Find the nearest location in under a millisecond, at any scale.

Architecture

Data in. Answers out in under a millisecond.

Ingest once. Orter distributes to worker nodes — self-hosted or in the cloud. Your software queries the nearest worker at wire speed.

0.4 ms p99 round-trip

120k+ lookups / sec / node

0.4 ms p99 round-trip

120k+ lookups / sec / node

Ingest

Push location data to the Management Portal.

Distribute

Replicated to worker nodes — self-hosted or cloud.

Query

Apps hit the nearest worker for instant geo lookups.

Capabilities

Everything you need, nothing you don’t.

Spatial search

Radius & nearest-neighbor, live

Query millions of points by radius or k-nearest in under a millisecond. Watch results land in real time across the map.

Performance

Sub-millisecond under load

A purpose-built lookup engine holds latency flat as throughput climbs. No tuning, no surprises.

One dashboard

Manage collections, worker nodes, and user access from a single modern interface.

Tiny footprint

Built in Rust. Blazingly fast while sipping minimal resources — runs anywhere, even a Raspberry Pi.

Any interface

REST, gRPC, and UNIX socket over one query engine. Use whatever fits your stack — no lock-in.

Developer experience

Drop into your stack.
no glue code.

Native SDKs for TypeScript, Python, and Rust. Raw HTTP and gRPC when you need it. UNIX socket for zero-latency local calls.

→Typed responses with full IntelliSense
→Built-in retries, timeouts, and connection pooling
→OpenAPI spec and Protobuf definitions on GitHub

curl -X POST https://worker.orter.dev/v1/query \
  -H "Authorization: Bearer $ORTER_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "lat": 52.5200,
    "lng": 13.4050,
    "radius_km": 5,
    "limit": 100
  }'

See it in action

From `docker pull` to first query in 60 seconds.

01 Deploy a worker

orter-worker — bash

→

02 Run your first query

first-query — bash

❯ curl -s localhost:8082/v1/query \

-d '{"lat":52.5200,"lng":13.4050,"radius_km":5}'

{

"results": [ /* 100 nearby points */ ],

"took_us": 412,

"total": 100

}

✅ 100 points · 0.41 ms — sub-millisecond, first try

Industries

Geospatial search, tuned to your industry.

Sub-millisecond nearest-neighbor search for AdTech, mobility, and logistics teams.

Win the bid before the budget runs out.

Programmatic auctions resolve in roughly ten milliseconds. Orter answers geo-targeting and geofence lookups in under one, so your DSP spends its budget on bid logic — not on a slow location query or a round-trip to a third-party API.

Resolve geo-targeting inside the ~10ms RTB bid window, with latency to spare for the rest of your bid.

Geofence and audience matching at full auction throughput — 120k+ lookups/sec per node.

Self-hosted: user location data never leaves your bid stack — no external geo API in the hot path.

<1ms

geo lookup

10ms

RTB budget

120k/s

bids / node

Interfaces

Talk with Orter how you want.

Pick whatever fits your stack — no lock-in, same query engine underneath.

REST

Plain HTTP and JSON. Works with any client, in any language — even raw curl.

gRPC

Typed, streaming, high-throughput. Generated clients for your favourite language.

UNIX socket

Zero network overhead for workers co-located with your app. Local-fast queries.

Create account, deploy worker, run first query — under three minutes.

FAQ

Questions, answered.