Scribe

Access provider-owned clinical documentation sessions, retrieve transcripts, generate notes and summaries, and review visit checklists and coding suggestions.

The Scribe API provides session-centric access to clinical documentation for authenticated providers. Each provider can list sessions they own, retrieve transcripts and coding suggestions, generate notes and summaries, and review the checklist supplied by the session's visit type.

Scribe has its own public OpenAPI schema and provider access boundary. Use the Scribe HTTP Reference for canonical operation contracts and Capture to Finalized Note for the complete workflow. Use the Scribe deployment base assigned to your integration, separately from Platform text sessions.

TypeScript SDK

Use @concurrence-hq/scribe-typescript-sdk 1.4.0 or later for Scribe REST helpers, streaming, and browser recording. Versions before 1.0.0 do not send the required session timezone, so their session creates fail with 422. Version 1.5.0 also types the carried value on carry_forward_available entries and the clinician-edit field dispositions that the service returns. The package is ESM-only and supports Node.js 20+ and browsers with the required capture APIs. Its integration guide explains the exported ScribeRecorder, separately provisioned provider access, and the backend route that obtains browser attach tickets. Keep provider credentials on your backend and enforce application access to the requested appointment or session.

npm i @concurrence-hq/scribe-typescript-sdk

import { ScribeClient, ScribeServerClient, ScribeStreamClient, ScribeRecorder } from '@concurrence-hq/scribe-typescript-sdk'

The SDK was previously published as @amigo-ai/scribe-typescript-sdk (last version 0.15.1). That package is deprecated and receives no further updates. Existing installs keep working at their pinned versions. To migrate, install @concurrence-hq/scribe-typescript-sdk and change the import path. Version 0.16.0 is the same API under the new name; 1.0.0 and later require timezone on session create and accept only the current visit types and note templates.

Authentication

Read operations require a valid provider token with scribe:sessions:read_own. Session creation, updates, lifecycle transitions, and Zoom controls require scribe:sessions:write. Artifact generation and finalization use their documented note-write scopes. Check the owning route rather than treating one scope as sufficient for every operation. The token must belong to a provider principal, and the provider's workspace must match the workspace in the request path. Send the token as a Bearer credential in the Authorization header.

Base Path

All Scribe endpoints are scoped under /v1/{workspace_id} and require the workspace identifier as a path parameter. Use the Scribe API base URL supplied for your deployment.

Resources

Resource	Description
Sessions	Create, inspect, and manage provider-owned clinical sessions, including Zoom capture
Artifacts	Retrieve transcripts and codes, follow note and summary generation, review checklist items, and finalize notes

Error Handling

The Scribe API returns structured error responses with a consistent envelope:

Field	Type	Description
code	string	Machine-readable error code (e.g., not_found, validation_error, unauthorized)
message	string	Human-readable error description
correlation_id	string (uuid)	Unique identifier for the request, useful for support inquiries
details	array	Optional list of field-level validation errors

Each detail object contains:

Field	Type	Description
field	string or null	The request field that caused the error
message	string	Description of the validation failure

Common Error Codes

HTTP Status	Code	Description
401	unauthorized	Bearer token is absent or invalid
403	forbidden	The principal lacks a required Scribe scope, provider binding, or workspace access
404	not_found	No provider-owned resource exists at this URL
409	Route-specific, such as conflict, version_conflict, or invalid_session_state	An empty transcript, a stale note version, or the session lifecycle prevents the requested operation
422	validation_error	Request validation failed
503	service_unavailable	The requested Scribe operation is temporarily unavailable

Temporary Maintenance and Reconnection

During maintenance, ordinary Scribe HTTP requests can return 503 with Retry-After. Honor the returned delay. Newly attaching streams can close with WebSocket code 1013 (try again later); an already attached stream is not terminated by this maintenance gate. Do not assume the installed SDK automatically retries every close code. Preserve the session ID, surface the interrupted state, and retry connection setup through the supported allocation and ticket flow after recovery.

Before repeating a note edit or finalization after an uncertain response, read the current note and version. A maintenance or transport error is not evidence that capture, generation, or a previous write completed.

Correlation IDs

Every response includes an X-Correlation-ID header. You can also send your own X-Correlation-ID header on requests - if valid, the API echoes it back; otherwise a new one is generated. Include the correlation ID when contacting support about a specific request.

OpenAPI Schema

The Scribe API publishes its OpenAPI schema at the /v1/openapi.json endpoint. This schema describes all available endpoints, request and response models, and validation constraints.