Deploy Agents on Chat.ai
Chat.ai Agents
Agents deployed on Chat.ai take advantage of the features provided by the Vital AI Agent Ecosystem
Agent features include:
- Deploy Agents privately, for your Teams, or for the Chat.ai membership
- Utilize Knowledge Graph and Documents
- Use commercial and open-source Large Language Models (LLMs) including GPT-4, Llama2, Claude, & Gemini
- Use any Agent development libraries such as LangChain, LlamaIndex, AutoGen, and Semantic Kernel
- Trigger actions such as sending outgoing emails
- Access external REST APIs
Agent Deployment
Agents may be deployed on Chat.ai in one of three ways:
- OpenGPTs — implementation based on LangChain's OpenGPTs library
- Client API — implementation using Websocket Client API
- Agent Container — implementation using a container to encapsulate the Agent implementation
In the OpenGPTs implementation, the Agent is specified in the Chat.ai web interface using prompts and may access resources within Chat.ai, such as the Knowledge Graph, as well as external REST APIs.
In the client-based implementation, requests to your Agent are sent to your implementation running on your infrastructure with your Agent sending replies back via the client.
In the container implementation, the Agent implementation is deployed via the container running within Chat.ai. The implementation has access to the Chat.ai resources, resources within the container including any code libraries, and external REST APIs.
Charge Usage Fees
Agents can be deployed and registered with the AgentShop allowing members of Chat.ai to install the Agent in the member's account, subject to fees.
Agents with usage fees:
- Fees can be based on a monthly subscription, per usage, or both, as specified by the Agent publisher
- Members may install or uninstall an Agent at any time
- Chat.ai collects fees from the membership and distributes usage fees to Agent publishers
- Chat.ai earns a commission on Agent usage fees. Fees incurred by the Agent, such as for LLM Model usage, are settled first.
- Chat.ai assists with discovery of Agents via the AgentShop and outreach to the membership.
Agent Features
The Vital AI Agent Ecosystem implements a robust framework for Agent development, deployment, and management.
Access Member Knowledge Graph
Agents can access the Knowledge Graph and Documents of the member, enabling personalization of the Agent behavior.
Agent Collaboration
Agents utilize AIMP, a standardized messaging protocol. With AIMP, Agents can communicate and collaborate with each other. This allows mixing and matching Agent capability.
Time, and the Passage of Time
Agents can be made aware of time and the passage of time, enabling more intelligent and fluid interactions.
Actions and Services
Agents can utilize abilities such as generating and executing code, browsing the Internet, sending and receiving email, and services that are available via REST API.
Contact Vital.ai
for your Agent Deployment
Contact Us
We are located at:
155 Water Street
Brooklyn, NY 11201