Suki is a developer platform for clinical documentation workflows that combines Ambient, Form filling, Dictation, and Patient Summary paths through APIs and SDKs. Public docs describe web, mobile, headless, and backend integration options, with feature access tied to partner configuration and some products in Beta. Mobile SDK support is iOS-only. [1][2][3][5]
Documented workflow areas
The docs present several workflow families rather than a single monolithic product path. Ambient documentation captures patient-provider conversation and turns it into structured clinical notes. Form filling converts conversation into structured form fields. Dictation returns real-time transcription for chart fields, editors, or scratchpads. Patient Summary generates pre-visit or encounter summaries from FHIR data. [1][3][4][5]
Suki also documents recommended E/M coding workflows. One path sends note sections and optional context to standalone E/M Coding APIs after an encounter. Another path ties coding recommendations to an ambient session and note retrieval flow. The documentation frames these as post-visit revenue cycle workflows, not as automatic billing decisions. [1][5]
Integration paths and UI control
The integration model varies by product and UI ownership. Web SDK ships hosted UI components for web healthcare apps. Headless Web SDK is described for React 18+ apps that need layout and branding control. Mobile SDK is headless for native mobile apps. Backend REST APIs and WebSockets support session management, audio transport, and result retrieval when the application owns more of the workflow. [2][5][6]
Form filling follows the same split. Suki documents REST APIs for custom UI or system integration, a standalone Form Filling SDK with hosted UI for JavaScript and React, and Web SDK support for teams already using the headed web package. For mobile, the docs state that Form filling on Mobile SDK is iOS only, and Android clients use Form filling APIs. [3][6][7]
One useful implementation detail
For teams building around Ambient or Form filling, the shared Partner WebSocket is a central transport layer for audio streaming, while REST APIs handle session lifecycle and retrieval. The docs describe webhooks as an alternative to continuous polling for asynchronous completion. That separation lets a developer choose synchronous polling, webhook-driven completion, or both, depending on internal architecture. [5][7]
Operational settings and context handling
The documentation describes session-based workflows that can accept context before or during capture. For Ambient, the session may include patient demographics, provider identity, specialty, visit type, encounter identifiers, and related clinical context. For Form filling, session context can be seeded or updated before audio is processed. For Patient Summary, the input is FHIR data stored in the Clinical Knowledge Graph. [1][4][5][7]
Output handling is also described at a systems level. After processing, applications can retrieve notes, structured form data, transcripts, or recommended codes and then write approved results into downstream EHR or clinical systems. The docs mention LOINC-based sections, diagnoses, orders when enabled, notification webhooks, and session completion or failure signals. [4][5][7]
Availability, authentication, and partner boundaries
Availability is not uniform across the catalog. The feature overview labels several products GA and several Beta, including Headless Web SDK, Dictation SDK, Form filling SDK, Patient Summary API, and E/M Coding APIs. The notes say Beta products may require partner configuration and that feature availability depends on partner configuration. [1][2][6]
Authentication is documented through public key sharing mechanisms. The quickstart lists Stored Secret, JWKS URL, Okta, and JWT Assertion. The API overview says Suki recommends OAuth 2.0 with JWT tokens and requires Authorization, partner_token, Content-Type, and Accept headers for API calls, while SDKs can manage authentication automatically after configuration. [3][5]
Implementation examples and boundaries
A practical implementation can begin with one workflow and add others later. The docs say workflows can be integrated independently. A team might start with Ambient note capture, then add dictation for inline edits, then add patient summaries before visits, or connect form filling for structured intake. That sequence is a documentation pattern, not a requirement. [1][3][4]
Hypothetical demonstration: a telehealth application could use Patient Summary before the appointment, Ambient during the visit, Form filling for intake fields, and Dictation for short clinician edits. The application could then request coding recommendations after the note is complete and send approved outputs to an EHR review queue. This is only an example flow, and the docs do not promise a specific clinical outcome.