How it Works
Actions
Adapt to user inputs seamlessly.
Ensure the LLM captures and addresses every nuance of the user's input.
Represent single shot & few shot prompting (with RAG) as well as fine-tuned models.
Action Graphs
Orchestrate multiple actions, whether in tandem or succession.
Create complex workflows with ease and precision.
Characters
Act as your application's intelligent thinking core.
Designed to examine problem statements, they possess the capability to reason, initiate actions, and gauge the outcomes.
Action Types
Include external connections like LLM models, diffusion models, web search, etc. and data retrieval service to build appropriate context.
User created action
For a particular action type (text to text, text to image etc), users can chat with them to iteratively get highly specific and tailored responses.
Action Graph
Acyclic graph composed of Actions and is triggered by user input.
Core Features
Retrieval Augmented Generation (RAG)
Self-Governing Agents
Capable of intelligent
independent action graph
creation and management
Enhanced Decision Autonomy
Accelerating logical decision making
Allowing systems to self-optimize
Execute with minimal human oversight
