SDK Integration

AxonFlow provides official SDKs for seamless integration with your applications.

Available SDKs

SDK	Package	Version	Status
Python	axonflow	6.9.0	Stable
TypeScript	@axonflow/sdk	6.2.0	Stable
Go	github.com/getaxonflow/axonflow-sdk-go/v8	6.0.0	Stable
Java	com.getaxonflow:axonflow-sdk	6.2.0	Stable
Rust	axonflow-sdk-rust	0.1.0	Preview

Requirements

SDK	Language Version	Package Manager
Python	Python 3.10+	pip
TypeScript	Node.js 14+	npm / yarn
Go	Go 1.21+	go modules
Java	Java 11+	Maven / Gradle
Rust	Rust 1.78+	Cargo

Integration Modes

AxonFlow supports two integration modes depending on your architecture:

Mode	Description	Best For
Proxy Mode	Route LLM calls through AxonFlow	Quick setup, full governance
Gateway Mode	Pre-flight policy checks, direct LLM calls	Low latency, existing infrastructure

Feature Support

The four stable SDKs (Python / TypeScript / Go / Java) ship the full feature set. The Rust SDK is in preview at v0.1.0 with the baseline shown below; subsequent Rust releases will fill in the rest.

Feature	Python	TypeScript	Go	Java	Rust (preview)
Proxy Mode	Yes	Yes	Yes	Yes	Yes
Gateway Mode (pre-check + audit)	Yes	Yes	Yes	Yes	Audit only
System Policy Management	Yes	Yes	Yes	Yes	No
Tenant Policy Management	Yes	Yes	Yes	Yes	No
Connector Management	Yes	Yes	Yes	Yes	List + query + install
Multi-Agent Planning (MAP)	Yes	Yes	Yes	Yes	Generate + execute + status + cancel
Audit Log Search	Yes	Yes	Yes	Yes	No
Health Check	Yes	Yes	Yes	Yes	No
failWorkflow()	Yes	Yes	Yes	Yes	No
HITL Queue API (Enterprise)	Yes	Yes	Yes	Yes	No
OpenAI Interceptor (invisible governance)	Yes	Yes	Yes	Yes	Yes

Not sure which to choose? See Choosing a Mode.

Quick Example (Python)

Async Usage (Recommended)

import asyncio
from axonflow import AxonFlow

async def main():
    async with AxonFlow(
        endpoint="https://your-axonflow.example.com",
        client_id="your-client-id",
        client_secret="your-client-secret"
    ) as client:
        # Execute a governed query
        response = await client.proxy_llm_call(
            user_token="user-jwt-token",
            query="What is AI governance?",
            request_type="chat"
        )
        print(response.data)

asyncio.run(main())

Sync Usage

from axonflow import AxonFlow

with AxonFlow.sync(
    endpoint="https://your-axonflow.example.com",
    client_id="your-client-id",
    client_secret="your-client-secret"
) as client:
    response = client.proxy_llm_call(
        user_token="user-jwt-token",
        query="What is AI governance?",
        request_type="chat"
    )
    print(response.data)

Gateway Mode (Lowest Latency)

Gateway Mode lets you make direct LLM calls while AxonFlow handles governance:

from axonflow import AxonFlow

async with AxonFlow(
    endpoint="https://your-axonflow.example.com",
    client_id="your-client-id",
    client_secret="your-client-secret"
) as client:
    # Pre-check: Get policy approval before LLM call
    context = await client.get_policy_approved_context(
        user_token="user-jwt-token",
        query="Analyze this data"
    )

    # Make your own LLM call directly
    llm_response = await your_llm_call(context.approved_data)

    # Audit: Log the response for compliance
    await client.audit_llm_call(
        context_id=context.context_id,
        response_summary=str(llm_response)[:200],
        provider="openai",
        model="gpt-4",
        token_usage=TokenUsage(prompt_tokens=50, completion_tokens=100, total_tokens=150),
        latency_ms=500
    )

Authentication

All SDKs support authentication via:

• Client ID + Secret - For server-to-server communication
• User Tokens - For per-user audit trails

See Authentication for details.

Next Steps

• Choose your SDK and follow the getting started guide
• Review Integration Modes to pick the right architecture
• Explore Framework Integration for LangChain, CrewAI, etc.
• See AI Agent Runtimes for OpenClaw, Anthropic Computer Use, and Claude Agent SDK