The REST query API (
/api/v1) is in early access and not yet open for external use. The code exists and the contract below reflects the current behavior. For external consumption now, use the OData feed.
Early access
Contract documented for planning. Do not build critical integrations on the REST API until it is released.
Returns paginated rows from a table in a published mart (*_gold/*_silver)
of the workspace. Read-only, scoped to the workspace.
Summary#
| Method · route | GET /api/v1/tables/{dataset}/{table}?limit=&offset= |
| Base | https://ingestia.io |
| Auth | API key (Authorization: Bearer ing_live_… or ?api_key=) |
| Role (RBAC) | data export (member+); data key = owner |
| Cost | charged by the actual bytes read in the datalake (see below) |
| Statuses | 200 · 400 · 401 · 403 · 429 · 402 · 502 · 503 |
Parameters#
| Parameter | Where | Default | Max | Notes |
|---|---|---|---|---|
dataset | route | — | — | must be {prefixo}_gold or {prefixo}_silver (prefixo = your workspace prefix); otherwise 403 |
table | route | — | — | only [A-Za-z0-9_] |
limit | query | 100 | 1000 | rows per page |
offset | query | 0 | — | starting offset |
Example#
curl -H "Authorization: Bearer ing_live_EXEMPLO" \
"https://ingestia.io/api/v1/tables/prd_aurora_gold/vendas?limit=100&offset=0"const url = "https://ingestia.io/api/v1/tables/prd_aurora_gold/vendas?limit=100&offset=0";
const res = await fetch(url, {
headers: { Authorization: "Bearer ing_live_EXEMPLO" },
});
const pagina = await res.json();import requests
res = requests.get(
"https://ingestia.io/api/v1/tables/prd_aurora_gold/vendas",
params={"limit": 100, "offset": 0},
headers={"Authorization": "Bearer ing_live_EXEMPLO"},
)
pagina = res.json()200 response#
{
"table": "vendas",
"rows": [ { "data": "2026-09-01", "total": 1234.56 } ],
"rowCount": 100,
"nextOffset": 100
}nextOffset is null when there are no more pages. The response includes the
x-ingestia-cost-brl header (projected cost of the call).
Cost and pagination#
Paginating a large table is the most expensive way to extract
LIMIT/OFFSET do not reduce the bytes read in the datalake — every page scans
the table. Billing uses the actual bytes processed. To pay less,
filter/aggregate with the OData feed (which pushes down
$select/$filter).
Errors#
| Code | When |
|---|---|
400 | invalid dataset/table name |
401 | missing/invalid key |
403 | dataset is not a published mart of the workspace / insufficient role |
429 · 402 | rate limit / overdue payment |
502 | failed to query the table in the datalake |
503 | limiter unavailable or datalake not provisioned |
Security#
- Published marts only (gold/silver); any other dataset →
403. - PII columns are not exposed.
- Billing and SQL are recorded in the workspace audit log.
Related#
Last reviewed: 2026-09-08.