Quickstart

Ceradela Storage in 5 minutes. Zero AWS setup on your side — it's a pure API. You send rows, you read them back.

1. Get a tenant name + token

Email hello@ceradela.com with a tenant name you'd like (e.g. acme). You'll receive back a 64-hex service token and the API base URL. Treat the token like a database password.

2. Push rows

Send a batch of rows and we encode them into BEBO (our columnar format), zstd-compress them, and store them under your tenant prefix. One HTTP call per batch, up to 50,000 rows per request. 

curl -X POST https://storage.../api/tenant/acme/ingest \
  -H "X-Service-Token: $TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "table": "orders",
    "rows": [
      {"id":1,"total_cents":12500,"created_at":"2026-04-01T00:00:00Z"},
      {"id":2,"total_cents":8900, "created_at":"2026-04-01T00:05:12Z"}
    ]
  }'

{
  "ok": true,
  "tenant": "acme",
  "table": "orders",
  "partition": "2026-04",
  "rows_stored": 2,
  "bytes_stored": 480,
  "storage_key": "acme/tables/orders/2026-04_ingest_1776563607442.cbebomth"
}

3. List your archives

curl https://storage.../api/tenant/acme/archives \
  -H "X-Service-Token: $TOKEN"

{
  "tenant": "acme",
  "rows": [{
    "storage_key": "acme/tables/orders/2026-04_ingest_1776563607442.cbebomth",
    "table_name": "orders",
    "partition": "2026-04",
    "row_count": 2,
    "byte_size": 480,
    "columns": "created_at,id,total_cents",
    "created_at": "2026-04-19T01:53:27Z"
  }],
  "total": 1,
  "page": 1,
  "limit": 50
}

4. Read rows back

curl "https://storage.../api/tenant/acme/cold/orders?partition=2026-04&page=1&limit=50" \
  -H "X-Service-Token: $TOKEN"

{
  "tenant": "acme",
  "table": "orders",
  "partition": "2026-04",
  "columns": ["created_at","id","total_cents"],
  "rows": [
    {"created_at":"2026-04-01T00:00:00Z","id":1,"total_cents":12500},
    {"created_at":"2026-04-01T00:05:12Z","id":2,"total_cents":8900}
  ],
  "total": 2,
  "page": 1,
  "limit": 50,
  "files": 1
}

5. Verify any archive locally (optional)

BEBO is an open format and the decoder is standalone. You can pull any archive down and inspect it yourself — this is the basis of our no-lock-in commitment. 

bebo verify --deep orders-2026-04.cbebomth
{ "decode_ok": true, "rows_decoded": 2, "sha256": "bd77..." }

bebo export orders-2026-04.cbebomth --to=csv

Mental model

  • You own the rows. We own the storage path.
  • No AWS setup on your side. We host S3, IAM, the Lambda, everything.
  • Per-tenant isolation. Your token only unlocks your prefix. Cross-tenant requests return 403 even with a valid token.
  • Atomic ingest. Each POST writes one BEBO file. No partial writes, no lost rows.
  • Monthly partitioning. Files bucket by the current UTC month automatically.
  • Exit is trivial. bebo export --to=parquet or download raw files. No lock-in.

What's next