SDK integrations

Pick an agent loop and see what xpander wires up for you

xpander supplies your agent’s identity (instructions, skills, knowledge bases, memory, deployment target) and delivers each task to your handler. In the SDK, skills are exposed as tools, the name agent loops use for them.

Agents created in Xpander Chat or through the API run on Claude Code, Codex or OpenCode by default. The SDK integrations on these pages are optional: you pick them when you want to run the LLM loop in your own process. The agent loop you choose is the library that does that: the thing with an Agent class, a function-calling interface, and an arun() method. xpander does not replace your agent loop; it supplies the agent definition and runs every skill call.

Selection lives in xpander_config.json, written by xpander agent new. The SDK reads the framework field from that file and wires up accordingly:

xpander_config.json

{
  "agent_id": "agt_01H...",
  "organization_id": "org_01H...",
  "api_key": "xpd_...",
  "framework": "agno"  // or: "open-ai-agents", "langchain", "strands-agents"
}

Right now, the SDK integrates with four agent loops:

Agno

OpenAI Agents SDK

LangChain

AWS Strands

Comparison

Capability	Agno	OpenAI Agents SDK	LangChain	AWS Strands
Pre-built skills from the catalog	Auto-included	agent.openai_agents_sdk_tools	agent.tools.functions	agent.strands_tools
Custom @register_tool functions	Auto-included	Auto-included	Auto-included	Auto-included
Instructions (role, goal, general)	Auto-attached	agent.instructions.full	agent.instructions.full	agent.instructions.full
Model + provider credentials	Auto-attached	agent.model_name (you build the client)	agent.model_name (you build the client)	agent.model_name (you build the client)
Knowledge-base retriever	Auto-attached	Manual via agent.knowledge_bases_retriever()	Manual	Manual
Session storage (Postgres)	Auto-wired	Manual	Manual	Manual
User and agent memory	Auto-wired	Manual	Manual	Manual
Context optimization (toon encoding, compaction)	Auto-wired	Not available	Not available	Not available
Guardrails (PII, prompt injection, moderation)	Auto-wired	Manual	Manual	Manual
Multi-agent teams	Auto-wired (AgnoTeam)	Manual	Manual	Manual
Session helpers (agent.get_user_sessions, etc.)	Yes	Raises NotImplementedError	Raises NotImplementedError	Raises NotImplementedError

A few takeaways worth calling out:

• Agno is the only path with automatic wiring. Backend.aget_args() dispatches on agent.framework. On the other three, you load the agent through Agents().aget(...) and read fields off it yourself.
• Pre-built skills and custom @register_tool functions reach every agent loop. What changes is the property name you read them from: agent.tools.functions for LangChain, agent.openai_agents_sdk_tools for the OpenAI Agents SDK, agent.strands_tools for AWS Strands.
• Memory, knowledge retrieval, and guardrails are Agno-only auto-wired. On the other agent loops, xpander gives you the data (agent.knowledge_bases_retriever(), session metadata, the agent’s memory config) but your code is responsible for plugging it into the loop.
• Agno-exclusive features: context optimization (toon encoding, runtime compaction), AgnoTeam-based multi-agent coordination, and the session helpers (agent.get_user_sessions, agent.get_session, agent.delete_session, which raise NotImplementedError outside Agno).

Model calls through the LLM Gateway

On the OpenAI Agents SDK, LangChain and AWS Strands you build the model client yourself. To send those calls through your organization’s LLM Gateway, so they run on the models your organization turned on, within your budgets, and show in Logs, give the client the gateway address as its base URL and a personal xpander API key as its API key. The OpenAI SDK takes the address with /v1; the Anthropic SDK takes it without. Your address and key are under Settings > AI models & keys in Xpander Chat; the steps are on Connect your tools.

Set XPANDER_GATEWAY_KEY to your personal xpander API key and replace https://llm-gw.xpander.ai with the Gateway address shown in AI models & keys when yours differs. Replace <model> with a model listed there.

OpenAI Python SDK, base URL with /v1:

import os
from openai import OpenAI

client = OpenAI(base_url="https://llm-gw.xpander.ai/v1", api_key=os.environ["XPANDER_GATEWAY_KEY"])
reply = client.chat.completions.create(model="<model>", messages=[{"role": "user", "content": "Reply with exactly: OK"}])
print(reply.choices[0].message.content)

Anthropic Python SDK, base URL without /v1:

import os
from anthropic import Anthropic

client = Anthropic(base_url="https://llm-gw.xpander.ai", api_key=os.environ["XPANDER_GATEWAY_KEY"])
reply = client.messages.create(model="<model>", max_tokens=64, messages=[{"role": "user", "content": "Reply with exactly: OK"}])
print(reply.content[0].text)

Each prints OK. The call shows under your name in Platform > AI governance > Logs (for admins) and in your own Settings > Usage. A 403 names its cause; see LLM Gateway errors.

How to choose

Within the SDK path, start with Agno unless you have a reason not to. Pick a non-Agno agent loop when:

• Existing investment. Your team already builds on it and switching cost is real. Example: a LangGraph workflow that’s been in production for six months.
• Loop-specific feature. You need something Agno doesn’t have. Example: LangGraph’s stateful multi-step workflows, the OpenAI Agents SDK’s Runner ergonomics, Strands’ AWS-native primitives.
• Embedded in an existing app. You’re adding xpander skills to a service that already runs one of these agent loops. Example: a FastAPI service that already imports agents.Runner.

You’re not locked in. The agent’s identity lives in xpander’s control plane, so you can swap agent loops later by editing framework in xpander_config.json and rewriting your handler.

Next steps

Agno

The recommended path. What Backend.aget_args() actually wires up.

OpenAI Agents SDK

Manual wiring with agent.openai_agents_sdk_tools.

LangChain + LangGraph

Manual wiring with agent.tools.functions and create_react_agent.

AWS Strands

Manual wiring with agent.strands_tools on AWS-native orchestration.

Core Concepts

The SDK class names mapped onto agents, tasks, threads, skills, and memory.

Quickstart

10-minute scaffold-to-deploy walkthrough on the Agno path.