floefox✳ Console
BIG ON DATA. LIGHT ON COST.

Large database.
Little price.

An extremely simple API.
Key/value. That's it.

Your data in object storage. Your hot reads in cache.
Built on SlateDB, with compute and storage separated.

Meet the $3 database Take a look inside

EARLY PREVIEW · NO CHARGES TODAY

COLD BY DESIGN. FOXY BY NATURE.An orange pixel fox keeping watch on an ice floe64° N / OBJECT STORAGE
01 SEPARATE COMPUTE & STORAGE02 CACHE THE HOT SET03 SIMPLE KEY/VALUE API

COMPUTE + STORAGE DISAGGREGATION

More data.
On your terms.

Compute serves requests and caches your working set. Object storage holds the durable database. Grow your dataset beyond local disk capacity, and size compute for your workload.

THE FLOEFOX WAY
ComputeMEMORY + DISK CACHE
Object storageTHE WHOLE DATASET
STORAGE GROWS. COMPUTE STAYS SEPARATE.

Writes wait for object-store durability before we acknowledge them. Cache misses fetch data from storage; cache hits take the shorter path.

HELLO, FLOEFOX 
curl -X PUT "$FLOEFOX_URL/v1/databases/$DB_ID/kv/aGVsbG8" \
  -H "Authorization: Bearer $FLOEFOX_API_KEY" \
  --data-binary 'hello, cold world'
YOUR KEY. YOUR VALUE. THAT'S THE API.

OBJECT STORAGE UNDERNEATH. ONE SIMPLE API ON TOP.

R2 today. Designed for S3-compatible object stores, with more backends to come. The same key/value API, whichever supported store holds your data.

Storage capacity and throughput scale differently. The preview uses one writer per database; read replicas and sharding are future work.

A LITTLE PRICE TO START

Start at $3.
Grow from there.

Start with the essentials. Explore how a bigger dataset or a larger cache could change the shape of your database.

THE $3 DATABASEPREVIEW
$3/ month

A starting point for your next big idea.

Database storage
10 GiB
Disk cache
256 MiB
Read target
100 / sec
Write target
10 / sec

Proposed launch targets for point reads and writes with 1 KiB values. Capacity and throughput are pending benchmarks; these are not current enforced limits.

Open the console

Private operator preview.
Public signup is coming. No billing today.

Find your footprint.

A LITTLE WHAT-IF MACHINE
10 GiB
1 GiB1,000 GiB

The full dataset in object storage.

256 MiB
256 MiB32 GiB

More room for frequently read data.

125 ms
20 ms200 ms

A planning target, rather than a performance guarantee.

ILLUSTRATIVE MONTHLY ESTIMATE
$3.00/ mo
Base plan$3.00
Extra storage$0.00
Extra cache$0.00
MODELED AVERAGE READ
118ms
Within your target
25% CACHE HITSMODEL ONLY
What goes into this estimate?

This is an illustrative planning model, not a quote or benchmark. It does not provision resources or change your database.

Pricing assumes a $3 base with 10 GiB storage and 256 MiB disk cache, then $0.10 per extra GiB of storage and $2 per extra GiB of cache each month. These rates are placeholders and do not model infrastructure cost or usage charges.

Latency assumes all reads are distributed uniformly across an active set equal to 10% of your database, with 20 ms cached reads and 150 ms cache misses. The modeled hit rate is cache size divided by active-set size, capped at 100%. The result is a weighted average, not p95 or p99. Real hit rates and latency depend on workload, network, cache state, and compaction; cold startup and writes are excluded.

THINK BIG. KEEP IT SIMPLE.

Give your data
a little breathing room.

Visit the console