Skip to main content
Version: v2.0
🚀 SDK v2 is here

v2 introduces a redesigned API with simplified agent creation, improved async execution, and first-class support for multi-agent workflows. If you're starting a new project, we recommend v2.

🛠️ Agent Builder Skill — build aiXplain agents in plain English from Claude Code, Cursor, and more.

aiXplain SDK

Build autonomous AI agents in a few lines of Python.

900+ models, tools, and integrations. Built-in memory. Run traces and dashboards out of the box. Granular runtime governance. One PAYG API key.

What you get

  • Single & multi-agent – Orchestrate subagents and let them delegate at runtime, no fixed flowcharts.
  • 900+ models, tools, MCPs – One PAYG key, swap providers without rewriting glue code.
  • Memory, traces, dashboards – Built in. Inspect every step, monitor every run.
  • Granular governance – Agents-in-the-loop policies enforced on every call.
  • Cloud or onprem – Serverless by default; sovereign deploys via Sovereignty.
Build and run an agent

Prerequisite: create a PAYG API key .

pip install aixplain
from aixplain import Aixplain

aix = Aixplain(api_key="YOUR_API_KEY")
agent = aix.Agent(
name="Research agent",
description="Answers questions with cited web findings.",
instructions="Use web search when needed and keep answers concise.",
tools=[aix.Tool.get("tavily/tavily-web-search/tavily")],
).save()

print(agent.run("What are the top 3 trending AI papers this week?").data.output)
Start building
Knowledge agents

Ground responses in your docs and knowledge bases with RAG workflows.

Data agents

Query operational data sources and return structured answers from live tables.

Custom logic agents

Add Python tools and business logic to execute domain-specific actions.

Integration agents

Connect to MCP servers and services like Slack, Salesforce, and more.

Team agents

Coordinate specialized agents for multi-step, multi-role execution.

Inspectors & guardrails

Enforce runtime quality and policy checks with inspectors and controls.

🧩 Marketplace MCP servers are now available

Extend 900+ models and tools with aiXplain-hosted MCP endpoints — accessible from any MCP client with a single PAYG API key.