Model ecosystem Companion · AI Companion directory

Hugging Face supplies the intelligence. DataInbox keeps the conditions.

Use Hugging Face Inference as an AI Companion for DataInbox. DataInbox controls context, permissions, routing, cost conditions, and output validation while Hugging Face supplies the AI capability.

Access to open-model inference and deployment tooling across a broad model ecosystem.
Approved inference endpoint, dedicated endpoint, or self-managed model serving.
Provider-independent task and result contract

Hugging Face Inference

External capability under a DataInbox policy boundary

Stable architecture, replaceable provider

DataInbox decides. The Companion performs.

A provider page describes where an external capability can fit. It does not move business logic, consent, permissions, or final authority into that provider.

Governed business message

DataInbox policy and context

Hugging Face Inference

The provider result returns to the Inbox as a structured message. Deterministic rules can validate it, reject it, route it for review, or allow a next action.

Provider fit

Where Hugging Face can fit

Access to open-model inference and deployment tooling across a broad model ecosystem.

Companion pattern: Model ecosystem Companion. DataInbox can select this provider directly for an eligible task.
Deployment boundary: Approved inference endpoint, dedicated endpoint, or self-managed model serving.

Stable capability profile

Broad model ecosystem
Open-model access
Deployment flexibility

Conditions to configure

Approved model and licence
Endpoint ownership
Security and capacity policy

Policy before inference

Control input, cost, provider, and output.

Data permission

Which fields and data classes may leave the Inbox.

Provider eligibility

Which direct provider, router, region, or private endpoint may be used.

Cost and service target

Budget ceiling, latency objective, availability, and fallback order.

Output contract

Required schema, evidence, confidence, validation, and review path.

Direct answers

Hugging Face AI Companion FAQ

What is the Hugging Face AI Companion?

It is a provider-specific implementation of the DataInbox Companion contract. DataInbox prepares a standardised AI task under business-controlled conditions, Hugging Face Inference performs the permitted capability, and the result returns to the Inbox as a governed business message.

Does using Hugging Face put business logic in the provider?

No. Business rules, permissions, context selection, routing policy, and result handling remain in DataInbox. The Companion supplies an external capability and returns a standardised result.

Can Hugging Face be replaced by another AI Companion?

Yes, when another Companion satisfies the same task and output contract. The replacement remains subject to company policy, connector availability, credentials, regional requirements, and validation.

Compare other AI Companion options

Provider choice remains a policy decision. Actual availability depends on the configured connector, credentials, region, contract, and deployment.

OpenAI

Direct AI Companion

Anthropic

Direct AI Companion

Google Cloud

Managed AI Companion

AWS

Managed multi-model Companion

Mistral AI

Direct or privately deployed AI Companion