• Memory: Enables long-term conversations by storing chat history and context in a database
• Knowledge: Provides LLMs with business context (RAG) by storing knowledge in Vector Databases
• Tools: Allows calling APIs, running queries, sending emails, and integrating LLMs with applications
• Monitoring: Tracks runs, tokens, feedback, quality and cost
• Evaluations: Automatically evaluates responses and sets alerts on quality issues

‍

• Building AI products and assistants
• Enhancing LLMs with additional capabilities like web search, database queries, etc.
• Creating domain-specific AI assistants with custom knowledge bases

‍

• Works with various LLMs including OpenAI, Anthropic, Cohere, and open source models
• Supports multiple databases and vector stores like Postgres, PgVector, Pinecone, LanceDb, Singlestore
• Can be served using Streamlit, FastAPI, or Django

‍