Sign Up
Login
Try for Free
Login
Outperforms general-purpose GPT-5.5, Claude Sonnet, and Gemini 3 Pro for healthcare applications while maintaining full HIPAA compliance. Purpose-built for therapy, clinical, and healthcare workflows.
// HIPAA-compliant chat completion POST https://api.bastiongpt.com/v1/ChatCompletion Content-Type: application/json Key: ab123456789012345678901234567890 { "messages": [{ "role": "user", "content": "What is a normal BP?" }], "max_tokens": 1000, "temperature": 0 }
This page outlines the necessary details to integrate with the BastionGPT API, including input parameters, output structure, pricing, limitations, and usage examples. To get started, sign up for BastionGPT Plus to explore our capabilities through the same API you'll receive access to. If you have additional questions or would like to apply for access, complete the access request form or schedule a consultation to review your use case and receive your API key.
Full compliance with healthcare regulations. Patient data stays secure with enterprise-grade encryption. Your data is never sold or used to train our AI models. Licensed for secure use with PHI and PII, with OpenAI having no access to your data.
Utilizes a pool of the leading AI models, fine-tuned for healthcare and mental health professionals. Expertly trained on thousands of clinical documentation standards and clinical workflows, such as SOAP notes, prior authorizations and appeal letters.
99.9% uptime SLA, 800 requests/minute, 2M tokens/minute. Built to support the next generation of powerful healthcare applications, with multi-region deployment in US, Canada (Quebec), and Australia for data residency compliance.
One POST request in, clinical-grade text out. Multi-turn conversations, your own system prompt, and a consistent response format with token usage on every call.
Server-Sent Events with OpenAI-compatible chunks, so responses render token by token in your UI and existing SSE client code just works.
Upload a PDF, image, or text file once and ask questions about it in any chat request. Documents are stored durably, scoped to your API key, and reusable across requests: upload once, reference the id as often as you need. Supported types: PDF, plain text, HTML, CSV, Markdown, PNG, JPEG, up to 10 MB.
Submit a recording, get back a transcript with speaker labels and timestamps. Medical transcription at $0.73 / hour, built for clinical accuracy on names, medications, and dosages. Common formats accepted (WAV, MP3, M4A, OGG, WEBM); conversion happens server-side.
Reduced consumer-style content filters so clinicians can discuss substance use, sexual health, end-of-life care, and other adult topics in a healthcare context, without dead ends.
Available in US, Canada (Quebec), and Australia for data residency compliance. Pick the region that matches your patients and regulatory footprint.
Start with a flat monthly subscription that includes credits, then pay-as-you-go beyond. No enterprise minimums.
Explore our capabilities over your 30-day developer trial through the same UI powered by the API you'll receive access to.
Schedule a 15-minute consultation or complete the access form below. Our technical and safety teams will review to confirm eligibility and ask any follow-up questions.
Get your key and full documentation shortly after the meeting, typically the same business day.
Make your first API call. Our team is on chat, email, or video whenever you'd like a hand.
For a full breakdown of costs before you request access, see API pricing and billing explained.
For an overview of what the API can do and common questions, see API capabilities and FAQ.
One base URL, an OpenAI-style request shape, and a consistent response format across chat, documents, and transcription. If you've integrated with mainstream LLM APIs before, you'll feel right at home.
One base URL for every request. Chat completions post to this endpoint.
https://api.bastiongpt.com/v1/ChatCompletion
Send messages, get a structured chat completion back. The messages array and max_tokens are required; temperature is optional and defaults to 0.
POST /v1/ChatCompletion HTTP/1.1 Host: api.bastiongpt.com Content-Type: application/json Key: ab123456789012345678901234567890 { "messages": [{"role": "user", "content": "What is a normal BP?"}], "max_tokens": 1000, "temperature": 0 }
Familiar choices/usage shape, with camelCase field names and token usage on every call.
{
"choices": [
{
"finishReason": "stop",
"message": {
"content": "A typical blood pressure range for a healthy adult is around 90/60 mmHg to 120/80 mmHg."
}
}
],
"id": "f1ad1c1a-4713-4d09-bcbe-7784e7434f2a",
"created": 1784062984,
"usage": {
"completionTokens": 26,
"promptTokens": 65,
"totalTokens": 91
}
}Complete reference documentation, a Postman collection, and an OpenAPI specification are provided with your API key.
Submit the form and our technical and safety teams will review for eligibility. We typically respond within one business day. Prefer a live conversation? Schedule a 15-minute consultation instead.
We prioritize the security and integrity of your data. Our API infrastructure is built with robust security measures, ensuring every interaction is protected using industry best practices.
Data in transit is protected by the latest, most secure encryption protocols. We exclusively support TLS 1.3 and TLS 1.2, disabling outdated and vulnerable protocols like SSLv3, TLS 1.0, and TLS 1.1.
Strong, authenticated cipher suites including AES-GCM (up to 256-bit) for high performance and the highest level of confidentiality and integrity for your API calls.
ECDHE key exchange ensures that even if a server's long-term private key were ever compromised, past communication sessions remain secure and cannot be decrypted.
Our API endpoint uses a certificate from GeoTrust / DigiCert, signed with a strong SHA-256 algorithm, verifying our server identity and establishing a trusted channel.
Configuration is actively hardened against known SSL/TLS vulnerabilities including BEAST, POODLE, Heartbleed, and ROBOT. We continuously monitor and update against emerging threats.
Our endpoint configuration consistently achieves a top 'A' rating from independent SSL/TLS assessment tools.
Start your trial, request access, or talk to our team. Whichever path fits, we'll have you making your first call within days, not quarters.
