delete_document
Delete Document
Permanently delete a document from Ragie. The document and all its chunks will be removed from the index.
delete_document
Permanently delete a document from Ragie. The document and all its chunks will be removed from the index.
retrieve_documents
Perform semantic search across indexed documents using a natural language query. Returns relevant text chunks suitable for use as LLM context. Supports metadata filtering, reranking for relevance, and limiting chunks per document for diversity.
update_document
Update an existing document's metadata, content from a URL, or raw text/JSON data. Metadata updates are applied without re-processing the document. Content updates replace the previous version and trigger re-indexing.
get_entities
Retrieve extracted entities either by instruction (all entities extracted by a specific instruction across documents) or by document (all entities from a specific document across all instructions). Use this to access structured data extracted from documents by entity extraction instructions.
manage_instructions
Create, list, update, or delete entity extraction instructions. Instructions define natural language prompts that Ragie automatically applies to documents to extract structured entities. Once created, an instruction is applied to all new and updated documents.
create_document
Ingest a new document into Ragie for indexing and retrieval. Supports creating documents from a **publicly accessible URL** or from **raw text/JSON data**. Use this to add content to your Ragie knowledge base. Documents go through processing states and become retrievable once indexed.
manage_partitions
List, create, inspect, or delete partitions. Partitions logically separate documents for multi-tenant applications or distinct knowledge bases. Retrievals can be scoped to a specific partition. Deleting a partition **irreversibly** removes all associated documents, connections, and instructions.
get_document
Retrieve detailed information about a specific document including its status, metadata, content, summary, and chunks. Use this to inspect a document's processing state, read its content, or access its chunks for analysis.
manage_connections
List, inspect, update, enable/disable, sync, or delete data source connections. Connections synchronize documents from external services like Google Drive, Notion, Confluence, and Salesforce. Use this to manage the lifecycle of your data source integrations.
list_documents
List documents in Ragie with optional metadata filtering and pagination. Returns documents sorted by creation date (newest first). Use this to browse your knowledge base, find documents by metadata, or paginate through all indexed content.
create_response
Generate an AI-powered answer using Ragie's deep-search agentic retrieval. The agent autonomously searches through your documents to find relevant information and synthesize a comprehensive answer with citations. Best for complex questions that require multi-hop reasoning across multiple documents.
Ingest, process, and semantically retrieve documents using RAG (Retrieval Augmented Generation). Upload files, URLs, or raw text for indexing across multimodal content including PDFs, images, audio, and video. Perform semantic search with metadata filtering, reranking, and chunk diversity controls. Extract structured entities from unstructured documents using natural language instructions. Manage data source connections to services like Google Drive, Notion, Confluence, and Salesforce for automatic document synchronization. Organize documents into partitions for multi-tenant isolation. Monitor document processing status and connection sync events via webhooks.
Common questions about connecting Ragie to AI agents with Metorial.