Dictaphone
MijnBureau supplies an installation of Dictaphone, also known as Assistant Transcripts. It is the transcription and meeting workflow application from LaSuite. The backend is a Django application with Celery workers, the frontend is a React single-page application, and audio extraction runs in dedicated transcoding and validation workers.
Configuration
To configure this solution, you can override the default settings for your environment. The
defaults are located in the folder helmfile/environments/default.
| Name | Description |
|---|---|
global.domain | The domain name of your MijnBureau instance |
global.hostname.dictaphone | The subdomain of the Dictaphone frontend |
global.hostname.dictaphone-minio | The subdomain used for pre-signed object storage URLs |
application.dictaphone.enabled | Enable Dictaphone |
application.dictaphone.namespace | The Kubernetes namespace name |
application.dictaphone.secretKey | Django secret key |
application.dictaphone.jwtSigningKey | JWT signing key |
application.dictaphone.audioExtractor.* | Internal audio worker tokens |
ai.dictaphone.* | Optional external transcription service |
authentication.client.dictaphone.* | OIDC client credentials |
database.dictaphone.* | PostgreSQL connection details |
cache.dictaphone.* | Valkey connection details |
objectstore.dictaphone.* | MinIO bucket and credentials |
tls.dictaphone.* | TLS settings |
pvc.dictaphone.* | Storage requirements for PostgreSQL, Valkey and MinIO |
resource.dictaphone.* | Resources for each component |
autoscaling.horizontal.dictaphone.* | Autoscaling for the backend, frontend and Celery worker |
container.dictaphone_* | Container image selection |
secret.dictaphone.* | Django superuser credentials |
AI transcription
Dictaphone sends transcription jobs to an external service compatible with the Meeting
Summary API. Set ai.dictaphone.serviceUrl to the service endpoint and provide
matching API keys through ai.dictaphone.serviceApiKey and
ai.dictaphone.webhookApiKey. When no AI service is configured, recordings and
transcripts can still be managed, but new transcription jobs cannot be processed.