Corti

Getting Started

Accessing the API

Getting an Access Token

Refreshing an Access Token

Using an Access Token

Interviews

Creates the interview in the database and returns it back to the client

Create interview

Fetching interview by ID for a prticular user

Get interview

Real-Time Interview Stream

Connecting to the WebSocket

Sending Events to the Server

Sending Data to the Server

Receiving Data from the Server

Languages

List all languages

Proved a lannguage code to get all details about a language

Get a specific language

Templates

List all templates

Get template

Transcription Stream

Corti Co-Pilot

APIs

Access Token

<p class="slate-p " >The Corti Co-Pilot API offers a comprehensive solution for integrating advanced speech-to-text, live note taking, and summary document generation into voice-based interactions. The overall flow of the API involves several key steps, from session initiation to real-time communication and session management.</p><h3 class="slate-h3 " >Overall Flow</h3><ol class="slate-ol " ><li class="slate-li " ><div style="position:relative"><span style="color:var(--tw-prose-bold)"><strong class="slate-bold">Initialization</strong></span>: The integration process begins with a POST request to <span style="color:var(--tw-prose-code)"><code class="slate-code"><strong class="slate-bold">/interview</strong></code></span>, which sets up a new session with unique authentication credentials and allows for the customization of AI features to suit specific medical consultation needs.</div></li><li class="slate-li " ><div style="position:relative"><span style="color:var(--tw-prose-bold)"><strong class="slate-bold">WebSocket Communication</strong></span>: A WebSocket connection is then established using the session ID and token, enabling the streaming of audio data from the consultation. This step is essential for facilitating real-time interaction between the client and the server.</div></li><li class="slate-li " ><div style="position:relative"><span style="color:var(--tw-prose-bold)"><strong class="slate-bold">Session Management</strong></span>: After the consultation, the session can be securely closed, or if necessary, it can be reopened within a 4-hour window to resume the interaction, ensuring flexibility and reliability in the documentation process.</div></li></ol><p class="slate-p " ></p><p class="slate-p " ></p>

The Corti Co-Pilot API offers a comprehensive solution for integrating advanced speech-to-text, live note taking, and summary document generation into voice-based interactions. The overall flow of the API involves several key steps, from session initiation to real-time communication and session management.

<p>Access to the Corti Co-Pilot API, particularly the development environment, is granted exclusively upon request to ensure that each integration meets the high standards of security and functionality required for medical applications. Interested customers are provided with a username and password after their request has been approved, which allows them to authenticate their applications and start integrating the API's capabilities.</p><p>Access to the API on</p>

Access to the Corti Co-Pilot API, particularly the development environment, is granted exclusively upon request to ensure that each integration meets the high standards of security and functionality required for medical applications. Interested customers are provided with a username and password after their request has been approved, which allows them to authenticate their applications and start integrating the API's capabilities.

Access to the API on all environments is governed by OAuth 2.0. This authentication protocol offers enhanced security measures, ensuring that access to patient data and medical documentation is securely managed and compliant with healthcare regulations. 

<p class="slate-p " >A POST request to <span style="color:var(--tw-prose-code)"><code class="slate-code"><strong class="slate-bold">/copilot/v1/interviews</strong></code></span> initiates a new session, generating a session ID and allowing configuration of AI settings, such as language preferences for transcription and documentation, as well as which template to use for the documentation.</p><p class="slate-p " >Participants are used for multi-channel streams. If you&apos;re only sending a single channel stream, you can leave the participants array empty, otherwise you&apos;d need to send a list of participants mapping each channel to the role of the participant in question. These participants will be referenced in incoming lines of transcript for multi-channel streams.</p><p class="slate-p " ></p><p class="slate-p " ></p>

 initiates a new session, generating a session ID and allowing configuration of AI settings, such as language preferences for transcription and documentation, as well as which template to use for the documentation.

Participants are used for multi-channel streams. If you're only sending a single channel stream, you can leave the participants array empty, otherwise you'd need to send a list of participants mapping each channel to the role of the participant in question. These participants will be referenced in incoming lines of transcript for multi-channel streams.

<p class="slate-p " >Creates the interview in the database and returns it back to the client</p><p class="slate-p " ></p><div class="slate-div"></div><p class="slate-p " ></p><p class="slate-p " ></p><div class="slate-div"></div><p class="slate-p " ></p><p class="slate-p " ></p>

language

documentation

participants

beta functionality leveraged in Corti Co-Pilot

specialty_id

Authorization

beta functionality leveraged by Corti Co-Pilot to generate and set titles

title

templateId

state

startedAt

endedAt

endsAt

transcript

factsGroup

facts

specialty

maxDurationInMinutes

detail

<p class="slate-p " >Fetching interview by ID for a particular user</p><p class="slate-p " ></p><div class="slate-div"></div><p class="slate-p " ></p><p class="slate-p " ></p><div class="slate-div"></div><p class="slate-p " ></p><p class="slate-p " ></p>

Fetching interview by ID for a particular user

interview_id

factGroups

free interviews are limited to 60 minutes in length

<table><colgroup></colgroup><tbody><tr><td><div><p><strong>Language</strong></p></div><div><div><div></div></div></div></td></tr></tbody></table>

type

event

data

Transcription Response

<p class="slate-p" placeholder="Type / for widgets">For cases where there is a need for real time transcription, for example voice commands or dictation, the system provides this by using a websocket connection</p><h2 class="slate-h2">Connecting to the WebSocket</h2><p class="slate-p" placeholder="Type / for widgets">With the interview ID and access token, a WebSocket connection can be established for streaming audio. This connection enables the transmission of audio data, receipt of real-time transcripts, and updates to documentation sections based on the audio input.</p><pre class="slate-CodeBlockElement slate-code_block"><div contenteditable="false"><div ><span aria-live="polite" aria-atomic="false" aria-relevant="additions text" ></span><div ><div ><div >Select...</div><div ><div  style="display:inline-block"><input type="text" autoCapitalize="none" autoComplete="off" autoCorrect="off" id="react-select-6-input" spellcheck="false" tabindex="0" value="" aria-autocomplete="list" style="box-sizing:content-box;width:1px;label:input;background:0;border:0;font-size:inherit;opacity:1;outline:0;padding:0;color:inherit"/><div style="position:absolute;top:0;left:0;visibility:hidden;height:0;overflow:scroll;white-space:pre"></div></div></div></div><div ><span ></span><div  aria-hidden="true"><svg height="20" width="20" viewBox="0 0 20 20" aria-hidden="true" focusable="false" ><path d="M4.516 7.548c0.436-0.446 1.043-0.481 1.576 0l3.908 3.747 3.908-3.747c0.533-0.481 1.141-0.446 1.574 0 0.436 0.445 0.408 1.197 0 1.615-0.406 0.418-4.695 4.502-4.695 4.502-0.217 0.223-0.502 0.335-0.787 0.335s-0.57-0.112-0.789-0.335c0 0-4.287-4.084-4.695-4.502s-0.436-1.17 0-1.615z"></path></svg></div></div></div></div></div><code class="keep-markup drop-tokens"><div class="slate-code_line">GET baseUrl/audio-bridge/v1/transcribe?token=Bearer%20$token</div></code></pre><h3 class="slate-h3">Sending data to the server</h3><p class="slate-p" placeholder="Type / for widgets">The only message type the client needs to send is an <code class="slate-code">audioBuffer</code>. The API currently supports webM and m4a audio encoding, and performs best when audio chunks are between 150-250ms (and no more than 1000ms) in length.</p><p class="slate-p" placeholder="Type / for widgets"><strong class="slate-bold">Example implementation:</strong></p><pre class="slate-CodeBlockElement slate-code_block"><div contenteditable="false"><div ><span aria-live="polite" aria-atomic="false" aria-relevant="additions text" ></span><div ><div ><div >Select...</div><div ><div  style="display:inline-block"><input type="text" autoCapitalize="none" autoComplete="off" autoCorrect="off" id="react-select-7-input" spellcheck="false" tabindex="0" value="" aria-autocomplete="list" style="box-sizing:content-box;width:1px;label:input;background:0;border:0;font-size:inherit;opacity:1;outline:0;padding:0;color:inherit"/><div style="position:absolute;top:0;left:0;visibility:hidden;height:0;overflow:scroll;white-space:pre"></div></div></div></div><div ><span ></span><div  aria-hidden="true"><svg height="20" width="20" viewBox="0 0 20 20" aria-hidden="true" focusable="false" ><path d="M4.516 7.548c0.436-0.446 1.043-0.481 1.576 0l3.908 3.747 3.908-3.747c0.533-0.481 1.141-0.446 1.574 0 0.436 0.445 0.408 1.197 0 1.615-0.406 0.418-4.695 4.502-4.695 4.502-0.217 0.223-0.502 0.335-0.787 0.335s-0.57-0.112-0.789-0.335c0 0-4.287-4.084-4.695-4.502s-0.436-1.17 0-1.615z"></path></svg></div></div></div></div></div><code class="language-typescript keep-markup drop-tokens"><div class="slate-code_line">// Setup the websocket
</div><div class="slate-code_line">const websocket = new WebSocket(&quot;https://api.beta-eu.corti.app/audio-bridge/v1/transcribe?token=Bearer%20TOKEN_VALUE&quot;);</div><div class="slate-code_line"></div><div class="slate-code_line">// Get the recording object
</div><div class="slate-code_line">navigator.mediaDevices
  .getUserMedia({ audio: true })
  .then((stream) =&gt; {</div><div class="slate-code_line">    // Create a MediaRecorder from the stream
    const recorder = new MediaRecorder(stream);
</div><div class="slate-code_line">    </div><div class="slate-code_line">    // Set the recorder to capture 250ms chunks of audio
    recorder.start(250);
</div><div class="slate-code_line"></div><div class="slate-code_line">    // When there is data available, send the data over the websocket
    recorder.addEventListener(&quot;dataavailable&quot;, (event) =&gt; {
      if (event.data.size &gt; 0 &amp;&amp; websocket.readyState == 1) {
        websocket.send(event.data);
      }
    });
});
</div><div class="slate-code_line"></div><div class="slate-code_line"></div></code></pre><h3 class="slate-h3">Receiving Transcriptions</h3><p class="slate-p" placeholder="Type / for widgets"></p><p class="slate-p" placeholder="Type / for widgets"></p><p class="slate-p" placeholder="Type / for widgets"></p><p class="slate-p" placeholder="Type / for widgets"></p><p class="slate-p" placeholder="Type / for widgets"></p><p class="slate-p" placeholder="Type / for widgets"></p>

For cases where there is a need for real time transcription, for example voice commands or dictation, the system provides this by using a websocket connection

With the interview ID and access token, a WebSocket connection can be established for streaming audio. This connection enables the transmission of audio data, receipt of real-time transcripts, and updates to documentation sections based on the audio input.

The only message type the client needs to send is an 

. The API currently supports webM and m4a audio encoding, and performs best when audio chunks are between 150-250ms (and no more than 1000ms) in length.