Which systems can it connect to?
CODEXIS AI is extensible through plugins, agents, skills and MCP servers.
CODEXIS AI is extensible through plugins, agents, skills and MCP servers. In practical terms, this means it can connect to systems that provide an API, a command line, a file interface or a dedicated integration service. The platform can work with plugin tools, local or marketplace skills, application components and connectors over MCP transports such as stdio, SSE or HTTP.
So typically you can connect, for example, an internal knowledge base, DMS, ERP, CRM, ticketing system, shared storage, a corporate API or another internal application, provided an appropriate connector exists or is supplied for it. This is not just passive data reading. Using these tools, an agent can perform workflow steps, run integration operations and return outputs back to the target system.
Example: a client may have a plugin that provides the tools lookupCustomer, createTicket and postReport, along with a synchronized folder of documents and a skill for an internal methodology. An automation then runs an agent every night that reads new files, evaluates them according to company rules, writes the result to the ticketing system and saves a report back to the synchronized folder.
6. What are the technical requirements and how do updates work?
How hardware requirements for CODEXIS AI are determined by workload, data and integrations, and how updates work.
8. Do I need to run CODEXIS AI on-premises to connect to cloud SharePoint?
Requirements for connecting CODEXIS AI to cloud-based SharePoint and when an on-premises deployment makes sense.