Point your existing SDK at https://fl1.trixycon.com and keep going. Multipart uploads, range reads, presigned URLs, versioning.
Loading service metrics…
# configure once aws configure set aws_access_key_id BSK7Q2XW9RMLE4TZ aws configure set aws_secret_access_key **************** # upload with automatic multipart aws --endpoint-url https://fl1.trixycon.com \ s3 cp ./dataset.tar.zst s3://backups/2026/ # range read, 64 MiB window aws --endpoint-url https://fl1.trixycon.com \ s3api get-object --bucket backups --key 2026/dataset.tar.zst \ --range bytes=0-67108863 part0.bin
# ~/.config/rclone/rclone.conf [basalt] type = s3 provider = Other endpoint = https://fl1.trixycon.com access_key_id = BSK7Q2XW9RMLE4TZ secret_access_key = **************** chunk_size = 64M upload_concurrency = 8 # sync a 4 TB tree, 8 streams rclone sync /srv/media basalt:media --transfers 8 --progress
import boto3
s3 = boto3.client(
"s3",
endpoint_url="https://fl1.trixycon.com",
aws_access_key_id="BSK7Q2XW9RMLE4TZ",
aws_secret_access_key="****************",
)
s3.upload_file(
"dataset.tar.zst", "backups", "2026/dataset.tar.zst",
Config=boto3.s3.transfer.TransferConfig(
multipart_chunksize=64 * 1024 * 1024,
max_concurrency=8,
),
)
Parts up to 5 GiB, 10 000 per object. Interrupted uploads resume from the last committed part.
Accept-Ranges: bytes on every object, so players and sync clients fetch only what they need.
SigV4 presigning with configurable TTL. Hand a link to a browser, keep your keys private.
Expire, transition to cold, or abort stale multipart uploads on a per-prefix schedule.
Sampled every 5 minutes from the edge fleet. Full history on the status page.
| Plan | Storage | Egress | Requests | From |
|---|---|---|---|---|
| Starter | 250 GB included | 1 TB / mo included | unmetered | $0 |
| Growth | $4 / TB / mo | $2 / TB | unmetered | $29 / mo |
| Scale | $3 / TB / mo | $1.40 / TB | unmetered | $249 / mo |
| Dedicated | reserved capacity | committed rate | unmetered | contact us |
Primary. Hydro-powered, 2 availability zones.
Read replica, 12 ms from Stockholm.
Edge cache, AMS-IX peering.