Documentation Index
Fetch the complete documentation index at: /llms.txt
Use this file to discover all available pages before exploring further.
Make your first Sage decision.
lv_live_...) in the dashboard under Decision model → API keys. Send it on every /decide and /decide/batch call:
Authorization: Bearer lv_live_...
SAGE_API_KEY.endpoint="https://sage.levanto.ai"
curl -s \
-H "Authorization: Bearer $SAGE_API_KEY" \
-H 'Content-Type: application/json' \
"$endpoint/decide" -d '{
"content": "Marketing email draft promotes '\''guaranteed 40% returns'\'' and describes the product as '\''risk-free'\'' for accredited investors.",
"question": {
"id": "needs_compliance_review",
"kind": "yesno",
"instructions": "Does this copy require compliance review before send?"
}
}'
{
"id": "needs_compliance_review",
"kind": "yesno",
"result": {
"answer": "yes",
"probability": 0.92
},
"meta": {
"model": "levanto-sage-v1.3",
"latency_ms": 97.4,
"reasoning": { "ran": false, "finished": null, "tokens": null }
}
}
content is what Sage evaluates (text, a list, or an image). question.kind picks the decision kind: yesno, choice, scale, sort or tags.
null to a human. probability is there if a workflow needs a stricter bar.
# request.json holds the request body from the step above
curl -s -H "Authorization: Bearer $SAGE_API_KEY" -H 'Content-Type: application/json' \
"$endpoint/decide" -d @request.json | jq -r '.result
| if .answer == null then "escalate"
elif .answer == "yes" then "send_to_compliance"
else "send" end'
https://sage.levanto.ai
| Method | Path | Key | Purpose |
|---|---|---|---|
| GET | /ready | No | 200 when ready, 503 while loading |
| POST | /decide | Yes | One decision |
| POST | /decide/batch | Yes | Many decisions |
| POST | /usage/estimate | Yes | Token count before you send |
Was this page helpful?