Lambent Probe API Documentation

Overview

The Lambent Probe API provides programmatic access to aphasia simulation, recovery tuning, and robustness testing. All endpoints require a valid API key.

Base URL: https://api.example.com/v1

Authentication: Bearer token in the Authorization header.

Rate Limit: 30 requests per minute per API key.

Model: patho-nlp-base v0.1.0-stub (rule-based engine)

Authentication

Include your API key in the Authorization header:

Authorization: Bearer lmb_your_api_key_here

POST /v1/simulate_aphasia

Simulates aphasic speech from normal text input.

Request Body

{
  "text": "Yesterday afternoon, I went to the grocery store to buy some fresh vegetables and fruit for dinner.",
  "aphasia_type": "broca",   // "broca" | "wernicke" | "conduction"
  "severity": 0.6            // 0.0 (mild) to 1.0 (severe)
}

Response

{
  "request_id": "550e8400-e29b-41d4-a716-446655440000",
  "model": "patho-nlp-base",
  "model_version": "v0.1.0-stub",
  "aphasia_type": "broca",
  "severity": 0.6,
  "original_text": "Yesterday afternoon...",
  "simulated_text": "Yesterday... store. Buy... vegetables.",
  "features": { "type": "broca", "agrammatism": true, "telegraphic": true },
  "processing_time_ms": 12
}

curl Example

curl -X POST https://api.example.com/v1/simulate_aphasia \
  -H "Authorization: Bearer lmb_your_key" \
  -H "Content-Type: application/json" \
  -d '{"text":"I went to the store to buy vegetables","aphasia_type":"broca","severity":0.7}'

Python Example

import requests

response = requests.post(
    "https://api.example.com/v1/simulate_aphasia",
    headers={"Authorization": "Bearer lmb_your_key"},
    json={
        "text": "I went to the store to buy vegetables",
        "aphasia_type": "broca",
        "severity": 0.7
    }
)
print(response.json())

POST /v1/recovery_tune

Simulates speech at a specific recovery stage from aphasia.

Request Body

{
  "text": "Yesterday afternoon, I went to the store.",
  "aphasia_type": "wernicke",
  "recovery_level": 0.5        // 0.0 (no recovery) to 1.0 (full recovery)
}

Response

{
  "request_id": "...",
  "model": "patho-nlp-base",
  "model_version": "v0.1.0-stub",
  "aphasia_type": "wernicke",
  "recovery_level": 0.5,
  "original_text": "Yesterday afternoon, I went to the store.",
  "recovered_text": "Yesterday afternoon, I went to the glorpicle.",
  "recovery_features": { "type": "wernicke", "recovery_level": 0.5 },
  "processing_time_ms": 8
}

POST /v1/robustness_test

Tests model robustness by applying various perturbations to input text.

Request Body

{
  "text": "The patient went to the hospital for treatment.",
  "perturbation_types": ["deletion", "substitution", "reorder"]
}

Response

{
  "request_id": "...",
  "model": "patho-nlp-base",
  "model_version": "v0.1.0-stub",
  "original_text": "The patient went to the hospital for treatment.",
  "perturbation_results": [
    { "perturbation_type": "deletion", "perturbed_text": "The went the for treatment.", "similarity_score": 0.72 },
    { "perturbation_type": "substitution", "perturbed_text": "The patienx went to the hospitax for treatment.", "similarity_score": 0.85 },
    { "perturbation_type": "reorder", "perturbed_text": "patient The to went hospital the treatment. for", "similarity_score": 0.68 }
  ],
  "processing_time_ms": 5
}

Error Codes

FHIR/HL7 Interoperability

Lambent provides a FHIR R4-compatible capability statement at:

GET https://api.example.com/api/integrations/fhir

This includes data mapping documentation for Patient and Observation resources. See our Data Protection Statement for details.