01
Receive any business message
Streams, APIs, databases, files, applications, devices, and MCP services enter through a controlled connection boundary.
Connectors
Connect anything. Govern every message.
DataInbox receives business messages from services, systems, streams, databases, files, devices, APIs, and MCP capabilities. Each message is validated and, where needed, translated into the standard contract expected by the receiving Inbox or third-party system.
- Any incoming message format
- MCP under Inbox conditions
- Schema validation and mapping
- Traceable delivery and results
Universal message boundary
Any source can participate. Every destination keeps its own contract.
If a service can send or receive through a protocol, adapter, API, or MCP capability, it can participate in a DataInbox operation. The Inbox between them determines which message is valid, which context may travel, and what the receiving party is allowed to get.
Streams and queuesEvents, topics, telemetry
Business systemsCRM, ERP, commerce, support
Data platformsDatabases and warehouses
Files and documentsStructured and unstructured input
APIs and webhooksPush, pull, subscriptions
MCP capabilitiesApproved tools and services
Governed message contract
- Validate identity and schema
- Apply purpose and permissions
- Translate fields and context
- Record delivery and result
Another InboxOnly permitted dossier context
Third-party systemsTheir required message contract
Companions and MCPCapabilities under conditions
Customer channelsConsistent governed responses
Ledgers and storesValidated records and evidence
People and approvalsDecisions with full context
The connector never becomes the authority. It transports a capability or message. The receiving Inbox remains responsible for validation, policy, translation, and evidence.
One connection model
From incoming payload to usable business message.
A connector moves data. DataInbox also establishes the context, contract, and conditions required to use that data safely.
02
Validate context and permission
The receiving Inbox checks schema, source, identity, purpose, permissions, and the conditions that apply to the message.
03
Map to the required contract
Mappings normalize fields and translate the message into the standard contract expected by the next Inbox or system.
04
Deliver and capture the result
The permitted message is delivered to its destination. Responses return as traceable business messages with current context.
Connector universe
Connected by protocol, not limited by a logo.
These categories show common connection patterns. Actual availability depends on the adapter, credentials, deployment, and receiving-system contract.
Event streams and message brokers7 common connection patterns
- Apache KafkaConsume topics as verified business events.
- Confluent CloudManaged Kafka streams with schema registry alignment.
- RedpandaKafka-compatible streams routed into the inbox.
- AWS KinesisStream ingestion with replay and ordering.
- Google Pub/SubPush or pull subscriptions, validated at ingest.
- Azure Event HubsHigh-throughput event capture and routing.
- NATS / RabbitMQSubject and queue based event routing.
Databases and warehouses7 common connection patterns
- PostgreSQLCDC via logical replication or polling.
- MySQL / MariaDBRow-level changes turned into events.
- MongoDBChange streams as inbox messages.
- SnowflakeQuery or stream tables into governed events.
- BigQueryScheduled or streaming export to inbox topics.
- DatabricksDelta table change feeds.
- RedshiftMaterialized streams or scheduled extracts.
SaaS and business apps8 common connection patterns
- StripePayments, refunds, disputes, payouts.
- SalesforceOpportunities, accounts, cases as events.
- HubSpotContacts, deals, marketing engagement.
- ShopifyOrders, customers, inventory, fulfillment.
- ZendeskTickets and conversation events.
- IntercomConversations, leads, product events.
- SAP / Oracle ERPDocuments and master data via adapters.
- NetSuiteTransactions and customer records.
Files and documents4 common connection patterns
- AWS S3Object events and structured file ingest.
- Google Drive / OneDriveDocument drops parsed to events.
- SFTP / FTPScheduled file pickup with validation.
- CSV / Parquet / JSONSchema-mapped into business events.
IoT and sensors3 common connection patterns
- MQTT brokersDevice telemetry as inbox events.
- OPC UAIndustrial machine signals.
- AWS IoT CoreDevice shadows and rule-routed payloads.
Webhooks and HTTP3 common connection patterns
- Inbound webhooksPer-source signed endpoints with replay protection.
- REST pollingScheduled pulls with delta tracking.
- GraphQL subscriptionsLive query results as events.
SDKs and bring your own5 common connection patterns
- JavaScript / TypeScript SDKClient and server event emission.
- Python SDKProducer for backend services and notebooks.
- Java / Go SDKsHigh-throughput producers.
- REST Event APIPOST any event to the inbox with a schema id.
- Datalayer / GTMStream browser events alongside analytics.
Connector FAQ
What data sources can DataInbox connect to?
DataInbox ingests from event streams (Kafka, Confluent, Redpanda, Kinesis, Pub/Sub, Event Hubs, NATS), databases (PostgreSQL, MySQL, MongoDB), warehouses (Snowflake, BigQuery, Databricks, Redshift), SaaS apps (Stripe, Salesforce, HubSpot, Shopify, Zendesk, Intercom, SAP, NetSuite), files (S3, Drive, SFTP, CSV, Parquet, JSON), IoT brokers (MQTT, OPC UA, AWS IoT Core), inbound webhooks, REST polling, GraphQL subscriptions, and SDKs for JavaScript, Python, Java, and Go.
How do I send a custom event to DataInbox?
Use the REST Event API, a webhook, an adapter, or an available SDK. The receiving Inbox validates the payload against its configured schema and can map it to the message contract required by the next system.
Are events validated at ingestion?
Yes. Every incoming message is validated against the configured schema before it becomes usable business context. Invalid messages receive a structured error and can be routed for review.
Can DataInbox run alongside Kafka or Confluent?
Yes. DataInbox consumes Kafka and Confluent topics as native sources and can also produce events back to them. It is not a replacement for the broker, it is the governance and routing layer above it.
Does DataInbox support GA4 or datalayer events?
Yes. Browser datalayer and GTM events can be streamed in parallel to GA4 so analytics and AI consume the same governed event stream without duplicate instrumentation.
Can MCP services connect to DataInbox?
Yes. An MCP service can provide an external capability under Inbox-defined conditions. DataInbox determines which context may be shared, which capability may be invoked, what result contract is required, and how the result returns as a governed business message.