Data Integration

Agents are only as useful as the data they can reach. Data integration is how you connect Hoziron agents to your existing systems — databases, REST APIs, file stores, and anything else your workflows need to read from or write to.

Integrations are configured through skills and credential stores, keeping connection details and secrets separate from agent logic. This means you can move agents between environments (dev, staging, production) without rewriting their competencies — only the integration configuration changes.

What you'll find here

GuideDescription
DatabasesConnect to PostgreSQL, MySQL, MSSQL, and other relational databases
APIsIntegrate with REST, GraphQL, and SOAP services
File systemsProcess documents from local paths, network shares, or S3-compatible storage
CredentialsManage secrets with vault integrations, environment variables, and rotation patterns

Common tasks

How data flows through Hoziron

Agents never touch raw credentials or connection strings directly. Skills mediate all external access, and the credential store handles secret retrieval at call time.

Security considerations

Data integrations are where your platform boundary meets the outside world. Keep in mind:

  • All credentials should go through the credential store — never hardcode secrets in competency configs
  • Use the principle of least privilege: grant agents only the database tables or API scopes they need
  • PII boundary enforcement applies to data flowing through integrations (see Workflows § Error handling)