Quickstart
Get from zero to your first compliance verdict in 5 minutes.
1. Get your API key
Go to Settings > API Keys in the dashboard and generate a new key.
Shell
export LITHRIM_API_KEY="lth_your_key_here"2. Create an agent
cURL
curl -X POST https://api.lithrim.com/agents \
-H "X-API-Key: $LITHRIM_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"name": "My Healthcare Agent",
"description": "Patient intake and scheduling bot"
}'3. Submit a transcript
cURL
curl -X POST https://api.lithrim.com/v1/analyze \
-H "X-API-Key: $LITHRIM_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"agent_id": "YOUR_AGENT_ID",
"transcript": "Agent: How can I help you?\nPatient: I need to refill my metformin.",
"artifacts": [
{
"type": "clinical_note",
"content": "Refill metformin 500mg BID. Identity verified.",
"target_system": "ehr"
}
]
}'4. Poll for results
Python
import time, requests
headers = {"X-API-Key": api_key}
while True:
resp = requests.get(f"{base_url}/v1/jobs/{job_id}", headers=headers)
data = resp.json()
if data["status"] in ("completed", "failed"):
break
time.sleep(5)
print(data["compliance"]["verdict"]) # "approve", "reject", or "needs_review"Verdict meanings
| Verdict | Meaning |
|---|---|
| approve | No violations. Safe to proceed. |
| reject | Violation found. Do not write to EHR. |
| needs_review | Uncertain. Queued for human review. |
| Artifact Verdict | Meaning |
|---|---|
| PASS | Faithful (score ≥ 0.9), no critical flags. |
| WARN | Mostly faithful (score ≥ 0.7), minor issues. |
| BLOCK | Hallucinated content or critical safety flags. Do not write to EHR. |