Overview
Sources & ingestion is how documents enter the K-AI Platform. K-AI does not require centralising the document estate — it addresses sources where they live, ingests them incrementally, preserves their native ACLs, and indexes them into the Semantic Document Layer.
What's in this section
Connectors — supported source types, one sub-page per connector. SharePoint, Confluence, Notion, Google Drive, Azure Blob Storage, Snowflake Stage, ServiceNow, generic HTTP, and the web crawler are documented in the section navigation. Start with the SharePoint connector; the others follow the same shape.
Document state machine — the lifecycle of a document from registration to indexed. See Document state machine.
Indexation pipeline — what runs between source and Semantic Document Layer. See Indexation pipeline.
Instance API — the machine-to-machine HTTP surface for indexation control. See Instance API overview.
Architectural placement
This section covers Layers 1 (Sources) and 2 (Ingestion & Indexing) of the 5-layer architecture. Layer 3 (Semantic Document Layer) is the Neural Semantic Graph. Layers 4 (Governance & Quality) and 5 (Exposure) are covered by K-AI Audit and K-AI MCP respectively.
Auth
The Instance API uses instance-id + api-key headers — see Authentication — Instance API keys.
Next steps
Pick a connector (e.g. SharePoint).
Start an indexation via the Orchestrator endpoint.
Monitor progress in the Documents endpoint.
Last updated