Skip to main content
Version: 2.0

Introduction

aiXplain Agents run on AgenticOS as autonomous runtime systems, not fixed workflow graphs. On each run, an agent can break a goal into multiple steps, select tools dynamically, call models and data sources, run code tools when needed, evaluate intermediate outputs, retry or switch strategy, and continue until completion criteria are met.

There are two ways to build and deploy them:

  • aiXplain Studio: visual builder for non-technical users and mixed teams.
  • SDK/API: technical path for developers who need code-level control.

How it fits together​

aiXplain agent execution flow on AgenticOS
  • aiXplain agents: your application-level agent behavior and toolset.
  • AgenticOS: runtime and control layer for execution, governance, memory, and observability.
  • Micro-agents: runtime control agents shown in the diagram: Mentalist, Orchestrator, Inspector, Bodyguard, and Responder. They coordinate planning, execution, validation, and guardrails on each run. See Inspectors.
  • Meta-agents: higher-level optimization/debug agents. The Evolver pattern is documented in v1 here: Evolver (v1).

Deployment flexibility​

  • You can deploy agents as Serverless or on-prem (private).
  • Serverless: managed deployment on aiXplain infrastructure (currently AWS in the US). See Serverless.
  • Private: deploy in your own environment on any cloud or server footprint, including bare metal, Kubernetes, and fully air-gapped setups, in any country/region. See Private.

Who this is for​

This documentation is for technical users who need to move from prototype to production without building orchestration, governance, and operations infrastructure from scratch.

Use aiXplain when you need:

  • Agent behavior that can adapt at runtime.
  • First-class integration with tools, data, and external services.
  • Access to 900+ AI models, tools, and integrations across vendors, with swappable models.
  • Production controls for security, cost, quality, and reliability.

What matters in production​

  • Reliability: retries, model/tool fallback chains, and runtime error handling.
  • Speed: fast prototyping, development, deployment, integration, and updates without rebuilding core orchestration.
  • Governance: inspectors, scoped API keys, RBAC, and rate limits.
  • Observability: full execution traces plus latency, usage, cost, and error metrics.
  • Sovereignty: keep control over data, models, and deployment boundaries. See Sovereignty.
  • Portability: managed cloud, VPC, and on-prem deployment options.
  • Extensibility: bring your own models, tools, data, code, and MCP servers.
  1. Start with Quick Start to run your first agent.
  2. Review AgenticOS for runtime architecture and production design.
  3. Add tools and integrations from Tools.
  4. Configure runtime policy with Inspectors.
  5. Choose a deployment mode: Serverless or Private.
  6. Configure operational controls with Workspaces, API Keys, and Credits & Billing.
  7. Use API Requests to integrate agents into your application stack.