API Documentation — RadRocket.AI

API Reference · v1.0

Build with
RadRocket.AI

Integrate AI-assisted radiology report generation, customizable report templates, report mistake checking, and medical-grade dictation into your applications.

6 endpointsREST-style APIToken authenticationRadiology-focused AI

Authentication

Every request requires an access_token. GET endpoints read it from the query string; POST endpoints read it from the request body. The server validates token expiry and the user's subscription plan.

Example API base URL

GET request example

curl -G "https://api.radrocket.ai/modality_list" \
  --data-urlencode "access_token=YOUR_ACCESS_TOKEN"

API tokens are credentials. Avoid exposing them in client-side code or logging GET request URLs containing access tokens.

Typical integration flow

Start with modality discovery, select a study and its template, then generate and review the report. Voice transcription is optional.

01 · Modality listChoose an imaging modality.

02 · Study listChoose an available study.

03 · Report templateLoad default report structure.

04 · Voice AI (optional)Transcribe dictated findings.

05 · Generate reportGenerate findings and impression.

06 · Mistake checkerReview before radiologist approval.

Endpoints

Request parameters and response field names are based on the uploaded backend definitions. Sample clinical values and IDs are illustrative.

List modalities

GET/modality_list

Retrieve available radiology imaging modalities and their identifiers.

Query parameters

Parameter	Type	Required	Description
`access_token`	string	Yes	API token
`sort`, `dir`	string	No	Declared optional query fields; sorting behavior not specified

Example request

cURL

curl -G "https://api.radrocket.ai/modality_list" \
  --data-urlencode "access_token=YOUR_ACCESS_TOKEN"

Example response

JSON · illustrative

[
  {"modality_id": 1, "modality_name": "Computed Tomography"},
  {"modality_id": 2, "modality_name": "Magnetic Resonance Imaging"}
]

List studies

GET/study_list

Return enabled default and user-owned studies for a selected modality, ordered by study name.

Query parameters

Parameter	Type	Required	Description
`access_token`	string	Yes	API token
`modality_id`	integer	Yes	ID from modality list
`sort`, `dir`	string	No	Declared optional query fields

Example request

cURL

curl -G "https://api.radrocket.ai/study_list" \
  --data-urlencode "access_token=YOUR_ACCESS_TOKEN" \
  --data-urlencode "modality_id=1"

Example response

JSON · illustrative

[
  {"study_id": 101, "template_name": "CT Brain Plain", "searchBy": "CT Head, NCCT Brain"},
  {"study_id": 102, "template_name": "CT Chest", "searchBy": "HRCT Chest, Thorax"}
]

Get report template

GET/report_template

Retrieve study heading, technique, default findings and impression, plus user-specific instructions.

Query parameters

Parameter	Type	Required	Description
`access_token`	string	Yes	API token
`study_id`	integer	Yes	Selected study identifier
`sort`, `dir`	string	No	Declared optional query fields

Example request

cURL

curl -G "https://api.radrocket.ai/report_template" \
  --data-urlencode "access_token=YOUR_ACCESS_TOKEN" \
  --data-urlencode "study_id=101"

Example response

JSON · illustrative

{
  "study_heading": "CT BRAIN PLAIN",
  "technique_text": "Non-contrast CT of the brain.",
  "study_custom_prompt": "",
  "findings_text": "No acute intracranial hemorrhage.",
  "impression_text": "No acute intracranial abnormality."
}

Some template fields may contain HTML formatting. Sanitize content before injecting it into a web page.

Generate radiology report

POST/generate_report

Generate a report from positive findings using a selected study and its reporting configuration.

Body parameters

Parameter	Type	Required	Description
`access_token`	string	Yes	API token
`study_id`	integer	Yes	Selected study
`positive_findings`	string	Yes	Observed radiological findings

Example request

cURL · form-encoded example

curl -X POST "https://api.radrocket.ai/generate_report" \
  -H "Content-Type: application/x-www-form-urlencoded" \
  --data-urlencode "access_token=YOUR_ACCESS_TOKEN" \
  --data-urlencode "study_id=101" \
  --data-urlencode "positive_findings=Small hypodense lesion in the right basal ganglia."

Example response

JSON · illustrative

{
  "study_heading_text": "CT BRAIN PLAIN",
  "technique_text": "Non-contrast CT of the brain.",
  "clinical_information_text": "",
  "findings_text": "Small hypodense lesion in the right basal ganglia.",
  "impression_text": "Small right basal ganglia hypodensity.",
  "comparison_text": ""
}

AI mistake checker

GET/mistake_checker

Analyze the submitted report text and return AI-generated feedback.

Query parameters

Parameter	Type	Required	Description
`access_token`	string	Yes	API token
`report_text`	string	Yes	Complete report text to analyze

Example request

cURL

curl -G "https://api.radrocket.ai/mistake_checker" \
  --data-urlencode "access_token=YOUR_ACCESS_TOKEN" \
  --data-urlencode "report_text=Findings: Left renal calculus. Impression: Right renal calculus."

Response field

Field	Type	Description
`mistake_checker_result`	string	Raw natural-language content returned by the configured model

Medical voice transcription

POST/voiceai

Convert radiologist dictation to text with radiology-oriented transcription instructions and normalization.

Body parameters

Parameter	Type	Required	Description
`access_token`	string	Yes	API token
`audio`	string	Yes	Audio payload; upstream format specified as WAV

Audio limits

Maximum upload size of 5,000,000 bytes (~5 MB). The upstream request uses format: "wav". Accepted encoding: base64

Example request

cURL · illustrative base64 form

curl -X POST "https://api.radrocket.ai/voiceai" \
  -H "Content-Type: application/x-www-form-urlencoded" \
  --data-urlencode "access_token=YOUR_ACCESS_TOKEN" \
  --data-urlencode "audio=BASE64_ENCODED_WAV"

Response fields

Field	Type	Description
`transcript`	string / null	Processed transcription; can be null
`transcript_copy`	string	Alternate branch output

Examples of normalization

Spoken	Output
Four by five	`4 x 5`
Five point one	`5.1`
Fifth	`5th`

Error handling

Status	Meaning	Suggested handling
401	Access token invalid or expired	Refresh or replace token
403	Plan not eligible / feature forbidden	Check subscription and permissions

Security & implementation notes

Transport: Use HTTPS and secure server-side storage for tokens. Avoid embedding credentials in distributed clients.

Patient data: Protect dictated audio and report text using appropriate access control, logging, retention and data-protection procedures.

Credential hygiene: The supplied Voice AI backend definition includes an embedded upstream credential. Rotate it and move it into secret management before sharing implementation files.

Publishing checklist: Confirm production route paths, POST content types, audio encoding, rate limits, billing behavior and standardized response/error schemas.