October 15: Support for Anthropic Claude models, Slack feeds and entity enrichment

New Features

  • 🔥 Graphlit now supports Anthropic Claude and Anthropic Claude Instant large language models.

  • 🔥 Graphlit now supports Slack feeds, and will ingest Slack messages and linked file attachments from a Slack channel. Note, this requires the creation of a Slack bot which has been added to the appropriate Slack channel.

  • 💡 Added support for entity enrichment to workflow object, which offers Diffbot, Wikipedia and Crunchbase enrichment of observed entities, such as Person, Organization and Place.

  • 💡 Added support for text extraction from images. When using Azure Image Analytics for entity extraction, Graphlit will extract and store any identified text which then becomes searchable.

  • Added embedFacets property to conversation strategy in specification object.

  • Added embedCitations property to conversation strategy in specification object. This makes content citations optional with the completed conversation message.

  • Added GraphQL mutations for multi-delete of entities, such as deleteCollections, deleteLabels, or deleteConversations.

  • Added GraphQL deleteAllConversations mutation to delete all conversations.

  • Added support for automatically adding ingested content to one or more collections, via ingestion stage of workflow object.

  • Added specification property to preparation workflow stage, which will be used to select the LLM for text summarization.

  • Expanded the properties for observed entities, such as Person, Organization or Product. Now supports a wider range of properties for entity enrichment.

Bugs Fixed

  • GPLA-1520: Unlimited conversation quota not assigned when upgrading project tier

  • GPLA-1285: Entity enrichment not firing event, which can be sent to actions

  • GPLA-1361: Web page left in ingested state, when URL not accessible.

Last updated