Worthic lets users create API clients for external tools, automations, and AI agents. Each client can be configured with a role, reporting-line scope, and specific capabilities, such as reading accounts and transactions, working with documents, generating reports, uploading files, proposing transactions, or writing transactions where allowed.
For AI agents, Worthic also includes an MCP server that exposes these capabilities as structured tools. An agent can list workspace context, search transactions, read accounts, find or download documents, upload documents, list available reports, and submit or propose transactions through the same permission model as the API.
This means users can connect external agent workflows without handing over unrestricted database access. The agent works through Worthic’s product layer, using the token’s approved scope and capabilities. Tokens can be reviewed, revoked, and audited from the workspace settings.