Platform Overview

Orlo Platform is where Orlo becomes an operational system, not just a set of packages.

It is the managed control plane teams use when AI is on a production decision path and the organization needs proof that the system was evaluated, deployed, governed, observed, and improved.

It helps organizations turn expert judgment into governed AI systems that improve with every decision.

When To Use Orlo Platform

Use Orlo Platform when:

  • multiple teams are adopting AI and need one shared operating model
  • model choice must be justified on task data
  • production responses need deterministic validation
  • retrieval quality and source attribution matter
  • feedback should improve future evaluations
  • agent tool actions need policy, approvals, limits, and trace evidence
  • risk, compliance, security, or audit need reviewable evidence
  • the system must run with tenant isolation, scoped access, and credential boundaries

Core Loop

text
Define -> Evaluate -> Deploy -> Govern -> Observe -> Improve
Loop Stage Platform Capability
Define Tasks, task versions, schemas, prompts, validation rules, owners
Evaluate Datasets, rubrics, candidate models, scoring, confidence intervals, budget controls
Deploy Deployment snapshots, activation, routing policies, fallback behavior
Govern Inference validation, retrieval attribution, routing metadata, agent step governance, approvals
Observe Mission Control, monitoring, inference logs, trace samples, trajectory scores, circuit breakers
Improve Feedback staging, correction review, trace promotion, dataset promotion, re-evaluation

Product Surfaces

Dashboard

The hosted dashboard is the operator surface for Orlo Platform.

It gives teams pages for:

The dashboard is not just a demo UI. It is the shared workspace for platform owners, domain leads, reviewers, and administrators.

Read Dashboard Surfaces for the user-facing role of each page.

Platform API

The Platform API is the machine interface for applications, integrations, SDKs, and hosted dashboard flows.

It covers:

  • task and dataset management
  • evaluations and deployments
  • task-native and OpenAI-compatible inference
  • feedback and promotion
  • document ingestion and retrieval-backed inference
  • credentials, models, adapters, org config, routing policies, and rubrics
  • monitoring and agent governance

Start with Platform API.

Gateway

The hosted gateway is the public boundary for Orlo Platform.

It owns:

  • hosted authentication and sessions
  • tenant and account context
  • API key issuance, rotation, revocation, scopes, and rate limits
  • RBAC and entitlement checks
  • dashboard BFF and proxy behavior
  • gateway audit and request logs
  • licensing for hosted and on-prem deployments

The control plane remains focused on Orlo product behavior: tasks, evaluations, deployments, inference, retrieval, feedback, credentials, and agent governance.

Workers

Longer-running work happens asynchronously:

  • evaluation jobs
  • document processing
  • embedding and retrieval preparation
  • scoring and statistics
  • log/archive workflows
  • monitoring and drift-related jobs

This keeps the runtime path bounded while still supporting deeper evidence and analysis.

What Platform Adds Beyond Open Core

Open Core gives reusable primitives. Platform adds the operating system around them:

  • tenant isolation and org-scoped persistence
  • gateway-backed access and API key management
  • task, dataset, evaluation, deployment, and feedback state
  • async workers and operational workflows
  • dashboard surfaces for operators and reviewers
  • deployment snapshots and inference logs
  • retrieval orchestration and attribution
  • agent governance persistence, approvals, limits, trace samples, and trajectory scoring
  • monitoring surfaces and evidence exports

Primary Integration Surface

The hosted API is served through https://api.useorlo.com.

Start with:

For responsible internal AI use and compliance evidence, also read Governance.