Platform API

Orlo Platform APIs are served from https://api.useorlo.com.

This is the application-facing API surface for the governed domain AI control plane.

It is not just a model endpoint. It covers the production decision path around domain AI:

• task definition and versioning
• evaluation and recommendation
• deployment and inference
• retrieval and document ingestion
• feedback and promotion
• agent governance and approvals

If your system needs to compare models before go-live, validate live outputs, retrieve from changing documents, or gate agent actions before they become consequential, this is the API surface that governs that path.

Auth model

Authentication requirements vary by deployment. These docs describe the Orlo-specific parts of the contract:

• trusted org-scoped request context via X-Orlo-Org-Id
• trusted task context via X-Orlo-Task-Id for OpenAI-compatible chat
• optional Idempotency-Key on mutating org-scoped requests

See Authentication for the details.

Public endpoint groups

• Health
• Tasks
• Datasets
• Evaluations
• Deployments
• Inference
• Feedback
• Documents and RAG
• Credentials
• Models
• Adapters
• Org Config
• Routing Policies
• Rubrics
• Monitoring
• Agent Governance

Machine-readable spec

• OpenAPI JSON

Platform behavior

• Orlo enforces tenant isolation with X-Orlo-Org-Id plus database row-level controls.
• Orlo keeps the main inference and governance surfaces asynchronous where heavy background work is required, especially for evaluation and PDF ingestion.
• Live inference can return routing metadata when the request is served through policy-based deployment selection.
• Agent-governance detail routes can return trajectory results, step-level path metrics, and reviewable trace samples.