How It Works

Connected Sources

A connected source uses a tracked sync job to move authorized content between MemorySync and an external system. Your application can trigger a job and inspect its public status without relying on worker details.

The customer-visible sync flow

01
Authorize
connection
Create a connection through the supported integration flow.
02
Configure
settings
Choose supported direction, frequency, and selection settings.
03
Trigger
job
Start a full or incremental job, or use the configured schedule.
04
Inspect
status
Store the returned job ID and read its public status and progress.
05
Recover
terminal
Handle partial or failed outcomes through documented controls.

Choose a job type

Job type	Use
`incremental`	Request changes since prior completed sync state. This is the default trigger value.
`full`	Request a complete sync for the connection’s authorized scope.

Trigger a sync job

trigger-sync.sh

curl --request POST https://api.memorysync.io/integrations/connections/42/sync \
  --header "Authorization: Bearer $MEMORYSYNC_ACCESS_TOKEN" \
  --header "Content-Type: application/json" \
  --data '{"job_type":"incremental"}'

response.json

{
  "id": 9001,
  "connection_id": 42,
  "job_type": "incremental",
  "status": "pending",
  "progress": {
    "total_items": 0,
    "processed_items": 0,
    "failed_items": 0
  },
  "started_at": null,
  "completed_at": null,
  "error_message": null,
  "created_at": "2026-07-31T12:00:00Z"
}

Inspect customer-visible status

Status	Meaning for your application
`pending`	The job exists and has not started running.
`running`	The job is in progress; read the progress object when present.
`completed`	The job completed.
`partial`	The job completed with some failures; inspect progress and error information.
`failed`	The job did not complete successfully; show a recoverable error state.
`cancelled`	The job was cancelled.

curl https://api.memorysync.io/integrations/sync-jobs/9001 \
  --header "Authorization: Bearer $MEMORYSYNC_ACCESS_TOKEN"

Use supported connection settings

Setting	Supported values
Frequency	`realtime`, `hourly`, `daily`, or `manual`
Direction	`import`, `export`, or `bidirectional`
Selection	Connection-specific include and exclude patterns

Available settings and behavior depend on the connector. Follow the connector page and keep source-system permissions authoritative.

Keep authorization and lifecycle explicit

Grant only the source-system permissions the connector needs.
Do not assume a source update or deletion changes an existing memory unless the connector contract says so.
Do not expose access tokens or connection credentials to browser code or logs.
Treat imported content as untrusted data and apply your application’s authorization before displaying it or sending it to a model.
Use job status and progress fields instead of depending on undocumented processing behavior.