Boostara Medical AI

Please enable JavaScript to use this site. Download the Boostara App on Google Play

Skip to main content Skip to main content
Boostara Innovations Logo
BOOSTARA MEDICAL AI
BY BOOSTARA INNOVATIONS
BIOMEDICAL CLOUD API & PROTOCOLS v4.8

Integrate Boostara Medical AI Into Any Clinical Workflow

Direct REST, WebSocket, and HL7/FHIR interfaces for electronic health records (EHR), picture archiving communication systems (PACS), DICOM viewers, telemetry monitors, and clinical workstations.

Base URL: https://api.boostarainnovations.co.ke/v1
Auth: Bearer In-App Token (AES-256)
FHIR Compatibility: R4 / R5 Strict
SECTION 01

Client Token Exchange & Zero-Trust Auth

Every API request to Boostara Medical AI requires a cryptographically signed Bearer token issued by your authenticated client device. Tokens are bound to your verified healthcare practitioner license (KMPDC / GMC / AMA) and the In-App Encrypted Vault session.

Requesting Header Specification

Attach your ephemeral session token in the Authorization header. Each token carries cryptographic proof of the client-side AES-256-GCM enclave.

// Standard HTTP Headers for Boostara Biomedical Endpoints
POST /v1/dicom/analyze HTTP/1.1
Host: api.boostarainnovations.co.ke
Authorization: Bearer bst_live_99f38e21aa47...
Content-Type: multipart/form-data
X-Boostara-Client-Id: clin_dr_sarah_0912
X-Model-Target: MODEL-9-OMNIPRESENT-RESEARCH
X-Vault-Storage-Proof: sha256_b3781fa09...
In-App Vault Guarantee: The raw image payload is processed strictly in volatile RAM across our secure inference cluster and automatically purged within 180ms of generating your structured report. The persistent copy resides solely on your client device.

Multi-Language SDK Initialization

Use our official high-performance client libraries for Python, Node.js, and Swift:

# Python 3.10+ SDK (pip install boostara-med)
import boostara_med as bst

client = bst.Client(
api_key="bst_live_99f38e21aa47...",
practitioner_id="KMPDC-A98144",
vault_path="/var/app/vault/in_app.db",
default_model=bst.Models.MODEL_9_OMNIPRESENT
)

# Verify real-time wallet token balance
balance = client.wallet.get_balance()
print(f"Available scans: {balance.scan_tokens} | Currency: {balance.currency}")
SECTION 02

Multi-Modal DICOM & Medical Imaging Inference

Execute full-resolution diagnostic screening over Chest X-Rays, Brain CT scans, Cardiac MRIs, Ultrasound cine loops, and Mammograms. Returns anatomical segmentations, bounding boxes, ACR BI-RADS / Fleischner risk scores, and ready-to-sign clinical reports.

POST /v1/dicom/analyze
Rate Limit: 600 req/min Latency: <180ms Model: Model 9 / Model 1

Request Body (Multipart)

Accepts .dcm, .nii, .png, .jpg
# cURL Invocation with Model 9
curl -X POST https://api.boostarainnovations.co.ke/v1/dicom/analyze \
-H "Authorization: Bearer bst_live_99f38e21aa47..." \
-F "file=@chest_pa_0428.dcm;type=application/dicom" \
-F "modality=CR" \
-F "body_part=CHEST" \
-F "clinical_indication=Pleuritic chest pain, persistent cough, fever 38.9C" \
-F "generate_gradcam=true" \
-F "hl7_format=FHIR_R4"
Query / Multipart Parameters:
file Binary File (Required) Raw DICOM or lossless medical image
modality String (Optional) CR, DX, CT, MR, US, MG (auto-detected if omitted)
generate_gradcam Boolean (Default: true) Returns normalized attention heatmaps

JSON Response (HTTP 200 OK)

Processing Time: 168ms
{
  "status": "SUCCESS",
  "inference_engine": "BOOSTARA-MODEL-9-OMNIPRESENT-RESEARCH",
  "execution_latency_ms": 168,
  "primary_modality": "CHEST_XRAY_PA",
  "image_quality": {
    "inspiration": "ADEQUATE_10_POSTERIOR_RIBS",
    "rotation": "NONE",
    "penetration": "OPTIMAL"
  },
  "findings": [
    {
      "condition": "Right Middle Lobe Consolidation",
      "confidence": 0.9942,
      "severity": "URGENT",
      "air_bronchograms": true,
      "silhouette_sign": "POSITIVE_RIGHT_HEART_BORDER",
      "bounding_box": [214, 480, 580, 720],
      "icd_10_code": "J18.9",
      "snomed_ct": "233604007"
    },
    {
      "condition": "Pneumothorax",
      "confidence": 0.0003,
      "status": "RULED_OUT"
    }
  ],
  "cardiothoracic_ratio": 0.44,
  "gradcam_heatmap_url": "blob:client_vault/heatmaps/cam_99214.png",
  "fhir_diagnostic_report_id": "urn:uuid:68a3e7b2-10f4-4112-98ab-301a2f9b8812",
  "vault_encryption_verified": true
}
SECTION 03

12-Lead ECG Voltage Telemetry & Arrhythmia Engine

Model 2 & Model 9 precision electrophysiology analyzer. Ingests raw voltage samples (500Hz / 1000Hz across I, II, III, aVR, aVL, aVF, V1V6) or high-resolution paper scans with automated grid calibration and perspective correction.

POST /v1/ecg/telemetry
Sampling: 500Hz / 12-Lead Latency: <92ms
Python Ingestion Example
import numpy as np
import requests

# 10-second 12-lead strip (12 x 5000 array)
raw_voltage_data = np.load("patient_telemetry_500hz.npy")

payload = {
"sampling_rate_hz": 500,
"leads": ["I", "II", "III", "aVR", "aVL", "aVF", "V1", "V2", "V3", "V4", "V5", "V6"],
"voltage_matrix": raw_voltage_data.tolist(),
"patient_age": 64,
"patient_sex": "M"
}

res = requests.post(
"https://api.boostarainnovations.co.ke/v1/ecg/telemetry",
headers={"Authorization": f"Bearer bst_live_EXAMPLE_KEY"},
json=payload
)
ECG Inference Report Output
{
  "status": "CRITICAL_EMERGENCY",
  "heart_rate_bpm": 118,
  "rhythm": "Sinus Tachycardia",
  "intervals_ms": {
    "PR": 142,
    "QRS": 96,
    "QT": 388,
    "QTc_Bazett": 448
  },
  "axis_degrees": 62,
  "critical_findings": [
    {
      "diagnosis": "Acute Anteroseptal STEMI",
      "confidence": 0.9981,
      "leads_involved": ["V1", "V2", "V3", "V4"],
      "elevation_mm": { "V2": 3.4, "V3": 4.1 },
      "reciprocal_depression": ["II", "III", "aVF"],
      "recommended_action": "IMMEDIATE CATH LAB ACTIVATION - CALL STAT CODE STEMI"
    }
  ]
}
SECTION 04

HL7/FHIR R4 Diagnostic Bundle Interoperability

Seamless bidirectional integration with Cerner, Epic, OpenMRS, Orthanc PACS, and dcm4chee. All Boostara findings are returned with standard LOINC, SNOMED CT, and ICD-10 codings mapped natively to DiagnosticReport and Observation resources.

FHIR R4 Resource: DiagnosticReport & Observation Bundle
IHE-RAD SOLE Validated
{
  "resourceType": "Bundle",
  "type": "transaction",
  "entry": [
    {
      "resource": {
        "resourceType": "DiagnosticReport",
        "status": "final",
        "category": [{
          "coding": [{
            "system": "http://terminology.hl7.org/CodeSystem/v2-0074",
            "code": "RAD",
            "display": "Radiology"
          }]
        }],
        "code": {
          "coding": [{
            "system": "http://loinc.org",
            "code": "24648-8",
            "display": "XR Chest PA and Lateral Views"
          }]
        },
        "conclusion": "Right Middle Lobe consolidation consistent with bacterial lobar pneumonia. No pleural effusion or pneumothorax.",
        "conclusionCode": [{
          "coding": [{
            "system": "http://snomed.info/sct",
            "code": "233604007",
            "display": "Pneumonia (disorder)"
          }]
        }]
      }
    }
  ]
}
SECTION 05

Ambient Clinical Dictation & SOAP Synthesis

Model 8 & Model 9 specialized medical speech-to-text. Ingests raw audio (WAV, FLAC, AAC, MP3) recorded at the bedside during patient rounds or clinical dictation. Generates formatted SOAP notes, prescription drafts, and follow-up orders within 500ms.

Clinical Audio Ingestion (POST /v1/clinical/dictation)

Multi-accent clinical speech recognition tuned across 85,000+ medical terms, pharmacopeia drugs, and regional clinician speech cadences (including African, Commonwealth, and North American accents).

# Send Bedside Recording via cURL
curl -X POST https://api.boostarainnovations.co.ke/v1/clinical/dictation \
-H "Authorization: Bearer bst_live_99f38e21aa47..." \
-F "audio=@ward_round_pt09.wav" \
-F "patient_context={\"age\":48,\"gender\":\"F\",\"bed\":\"ICU-3\"}" \
-F "output_format=SOAP_NOTE"

Synthesized SOAP Output

[SUBJECTIVE]
Patient is a 48-year-old female reporting 4 days of progressive shortness of breath, productive cough with rust-colored sputum, and right-sided pleuritic pain. Denies hemoptysis.

[OBJECTIVE]
Vitals: BP 118/74 | HR 102 | RR 24 | Temp 38.8C | SpO2 91% on RA.
Lungs: Crackles and bronchial breath sounds over right middle lung field. Dullness to percussion at right mid-zone.

[ASSESSMENT]
1. Community-Acquired Pneumonia (CAP) - Right Middle Lobe (ICD-10: J13)
2. Mild Type 1 Hypoxemic Respiratory Failure (ICD-10: J96.01)

[PLAN]
1. Start Ceftriaxone 1g IV daily + Azithromycin 500mg PO daily.
2. Low-flow O2 via nasal cannula at 2 L/min to target SpO2 > 94%.
3. Serial bedside lung ultrasound at 24 hours.
SECTION 06

WebSocket Real-Time Ultrasound Cine Streaming

Designed for handheld POCUS probes (Butterfly iQ, Clarius, GE Vscan). Streams 30 frames per second over persistent TLS WebSockets with frame-by-frame cardiac ejection fraction (Simpson's biplane), B-line quantification, and vascular needle guidance.

wss://api.boostarainnovations.co.ke/v1/stream/ultrasound
POCUS STREAM ENGINE
// TypeScript / Node.js POCUS Handheld Client Stream
const socket = new WebSocket("wss://api.boostarainnovations.co.ke/v1/stream/ultrasound");
socket.onopen = () => {
socket.send(JSON.stringify({
action: "INIT_STREAM",
probe_type: "PHASED_ARRAY_CARDIAC",
token: "bst_live_99f38e21aa47..."
}));
};

socket.onmessage = (event) => {
const frameAnalysis = JSON.parse(event.data);
console.log("Instant Ejection Fraction:", frameAnalysis.ejection_fraction_pct); // e.g. 56.4%
console.log("B-Lines Detected:", frameAnalysis.b_line_count); // e.g. 1 (Normal)
};
SECTION 07

Diagnostic Safety Flags & HTTP Error Codes

Boostara Medical AI prioritizes patient safety above all else. When an image has severe artifacts, patient motion, or sub-diagnostic quality, the model emits explicit safety flags rather than hallucinating clinical findings.

HTTP Status Safety Code Clinical Rationale Recommended Protocol
422 Unprocessable SUBOPTIMAL_INSPIRATION Less than 8 posterior ribs visible on PA chest film; risk of false cardiomegaly. Advise radiographer to repeat image on deep inspiratory hold.
422 Unprocessable MOTION_BLUR_ARTIFACT High-frequency phase degradation prevents detection of micro-fractures or fine infiltrates. Re-acquire view with patient breath-hold or limb immobilization.
200 OK + STAT Flag CRITICAL_FINDING_STAT Imminent life threat detected: Tension pneumothorax, massive PE, aortic dissection, acute STEMI. Automatic push notification dispatched to on-call clinician device within 1.2s.
402 Payment Required WALLET_TOKENS_EXHAUSTED Clinician account balance depleted below scan token threshold. Trigger instant M-Pesa STK push via In-App Wallet for immediate token reload.

Start Integrating Boostara Medical AI Today

Download the official Boostara Client App to obtain your cryptographic practitioner token, review token economics, and test live DICOM inference.

Chat on WhatsApp

Clinical Disclaimer: Boostara Medical AI products are AI-assisted decision support tools intended for use by qualified licensed healthcare professionals. They are not a substitute for professional medical advice, clinical diagnosis, or treatment. All AI-generated findings must be reviewed and validated by a licensed clinician. Not cleared as a medical device. For support: support@boostarainnovations.co.ke