AI operating system
for your server.
River runs your databases, applications, files, and analytics on hardware you own, containers in any language, and an AI that runs the whole platform with you.
Trusted by data teams at · A closer Look · Lobby Game Studio · Kitawi Systems · Stella
01–14 · The platform
One operating system, every layer of your stack.
Databases, runtimes, files, dashboards, and AI — running on hardware you own. Open any capability to see it working.
Run databases and connected services from one control plane.
$ river provision --postgres
✓ instance reserved · 2 vCPU · 4 GB
✓ pg1 running — ready in 48s
Query across every backend without copying your data.
SELECT u.name, o.total JOIN mongo1.orders o ON u.id = o.user_id -- pushdown: pg1 · mongo1
One assistant that runs the whole platform — code, servers, schemas, and queries.
Run backend applications in any language on your own server.
Browse, filter, and edit real tables without writing a query.
Move from exploration to reproducible analysis in one workspace.
Turn live queries into dashboards your team can trust.
Evolve schemas like code, with history and confidence.
Automate recurring work with visible runs and outcomes.
Bring files into analysis — and put sites on the internet.
See the machine, the databases, and the queries that run on them.
Own the box, and control exactly who reaches what.
Typed clients for the languages your team already ships.
Every capability, on one predictable HTTP surface.
The engine underneath
One query, every backend — data stays where it lives.
Every River capability — Lily, notebooks, BI tiles, jobs, the table editor — reads through the same federated engine. That is what makes an operating system out of a stack of tools.
-- join Postgres ⇄ MongoDB without copying
SELECT u.name, o.total
FROM pg1.users AS u
JOIN mongo1.orders AS o
ON u.id = o.user_id
WHERE o.total > 400
ORDER BY o.total DESC
LIMIT 5;✓ pushdown complete · pg1 (24 ms) · mongo1 (216 ms) · merged 5 rows
Pushdown to source engines
River rewrites your query and pushes the heavy work to Postgres, MySQL, MongoDB, and more — so you read data where it already lives.
Zero-copy cross-database joins
Join a Postgres table to a MongoDB collection without moving either into a central warehouse. No ETL, no drift, no stale copies.
One consistent SQL dialect
Write a single query and let River translate it per backend. Your team stops juggling six dialects and six drivers.
Pricing
Your data, on hardware you own.
Choose the dedicated environment that fits your current. Open any option for a complete infrastructure breakdown and checkout estimate.
Starter
Cloud VPS 4
A focused home for side projects and proofs of concept.
| CPU | 4 vCPU |
| Memory | 8 GB RAM |
| Storage | 100 GB SSD |
| Snapshots | 1 |
Growth
Cloud VPS 8
Production capacity for a team building on one dependable instance.
| CPU | 8 vCPU |
| Memory | 24 GB RAM |
| Storage | 300 GB SSD |
| Snapshots | 3 |
Scale
Cloud VPS 18
More room for federated backends, notebooks, and BI in one place.
| CPU | 18 vCPU |
| Memory | 96 GB RAM |
| Storage | 600 GB SSD |
| Snapshots | 3 |
Pro
AMD Ryzen 12 Cores
Dedicated bare metal for latency-sensitive workloads.
| CPU | 12 × 3.70 GHz Ryzen 9 7900 |
| Memory | 128 GB DDR5 |
| Storage | 2 TB NVMe |
| Network | 1 Gbit/s |
Enterprise
AMD Genoa 24 Cores
High-concurrency infrastructure for serious production analytics.
| CPU | 24 × 2.50 GHz EPYC 9224 |
| Memory | 128 GB REG ECC (up to 768 GB) |
| Storage | 2 × 1 TB SSD |
| Network | 1 Gbit/s |
Custom
AMD Turin 32–64 Cores
A tailored River environment for multi-node, HA, and compliance needs.
| CPU | AMD EPYC Turin 32 or 64 core |
| Memory | 128 GB to 1.15 TB |
| Storage | 2 × 1 TB NVMe and custom |
| Network | Up to 10 Gbit/s |
Annual billing can save approximately two months. Setup and migration are available on Scale and above.
FAQ
Questions, answered.
River is an AI operating system for your server. It runs your databases, containers, notebooks, files, dashboards, and jobs on hardware you own — with federated SQL across every source, containers in any language, SDKs in seven languages, and a complete HTTP API.
Own the box · own the data
Spin up River on hardware you own.
Databases, containers, dashboards, files, and an AI that ships queries — one operating system, no shared tenancy, no data hand-off.