Integrate TinyFish with Jinba

Yes. Jinba is designed specifically for regulated enterprises that require on-premise or private-cloud deployment. When you integrate TinyFish through Jinba Flow, workflows can be hosted entirely within your own infrastructure. Jinba supports private model hosting via AWS Bedrock, Azure AI, or self-hosted models, ensuring that sensitive documents never leave your controlled environment. This is a core requirement for banks, insurance companies, and other enterprises operating in air-gapped or strictly governed deployment contexts.

Not if you configure private model hosting. Jinba supports private deployment options including AWS Bedrock, Azure AI, and self-hosted models, so your document data and workflow inputs remain within your own infrastructure. This is a deliberate architectural choice for regulated industries — Jinba's team can provide engineering support to configure your specific hosting environment and ensure data does not transit through public AI endpoints.

Jinba Flow wraps TinyFish AI capabilities in a deterministic execution layer where 80% of workflow logic is rule-based. Every workflow execution is logged in full audit trails. Combined with role-based permissions (RBAC), SSO, and Active Directory integration, compliance teams can trace every action back to the user and workflow version that triggered it. Version control and feature flags provide additional governance for production rollouts.

Yes. Workflows built with TinyFish in Jinba Flow are published as shared team automations accessible through Jinba App, a conversational execution interface with auto-generated input forms. Non-technical operations staff such as KYC analysts, compliance officers, and loan processors can invoke approved workflows without accessing the underlying builder. Access is controlled by RBAC, so only authorized roles can run specific workflows, reducing operational risk.

Jinba Flow is designed to reduce workflow build time significantly compared to traditional implementation approaches. Teams can describe a workflow in natural language using chat-to-flow generation, refine it in the visual editor, test it with real data, and publish it as an API or batch process. This approach is intended to compress timelines from months to days, without the need for large consultant-driven projects or bespoke engineering work.