AI Agent Integrations

Connect OpenClaw, ZeroClaw, Nanobot, and other Claw-family AI agents to Swan Inference

Swan Inference provides an OpenAI-compatible API that works with all Claw-family AI agent tools. This guide covers how to connect each tool to Swan Inference for decentralized AI inference.

Prerequisites

A Swan Inference API key (sk-swan-*) — sign up here
The Claw tool of your choice installed

Quick Start

All Claw-family tools support custom OpenAI-compatible endpoints. The core configuration is the same across all tools:

Setting

Value

Base URL

https://api.swanchain.io/v1

API Key

sk-swan-YOUR-API-KEY

Model

Any model from inference.swanchain.io/models

Popular models available on Swan Inference:

Model

Tool-Specific Setup

OpenClaw

The most popular AI assistant in the Claw family. Supports 25+ messaging platforms.

Edit your OpenClaw configuration to add Swan Inference as a provider:

{
  "models": {
    "providers": {
      "swan": {
        "api": "openai-completions",
        "baseUrl": "https://api.swanchain.io/v1",
        "apiKey": "sk-swan-YOUR-API-KEY",
        "models": [
          "zai-org/GLM-4.7-Flash",
          "Qwen/Qwen2.5-7B-Instruct"
        ]
      }
    },
    "default": "swan:zai-org/GLM-4.7-Flash"
  }
}

export OPENAI_API_BASE=https://api.swanchain.io/v1
export OPENAI_API_KEY=sk-swan-YOUR-API-KEY
openclaw

ZeroClaw (Rust)

Ultra-lightweight (3.4MB binary, <10ms startup). Best for quick testing.

[provider]
name = "openai-compatible"
base_url = "https://api.swanchain.io/v1"
api_key = "sk-swan-YOUR-API-KEY"
model = "zai-org/GLM-4.7-Flash"

zeroclaw --provider "custom:https://api.swanchain.io/v1" \
         --api-key "sk-swan-YOUR-API-KEY" \
         --model "zai-org/GLM-4.7-Flash"

PicoClaw (Go)

Lightweight Go binary (<8MB). Designed for edge and ARM devices.

# config.yaml
provider:
  type: openai-compatible
  base_url: https://api.swanchain.io/v1
  api_key: sk-swan-YOUR-API-KEY
  model: zai-org/GLM-4.7-Flash

Nanobot

Python-native, installable via pip. Best for Python developers.

pip install nanobot-ai

{
  "provider": "custom",
  "apiBase": "https://api.swanchain.io/v1",
  "apiKey": "sk-swan-YOUR-API-KEY",
  "model": "zai-org/GLM-4.7-Flash"
}

from nanobot import Agent

agent = Agent(
    provider="custom",
    api_base="https://api.swanchain.io/v1",
    api_key="sk-swan-YOUR-API-KEY",
    model="zai-org/GLM-4.7-Flash",
)

response = agent.chat("Explain quantum computing in simple terms")
print(response)

NanoClaw (TypeScript)

Container-per-session isolation. Good for multi-tenant deployments.

{
  "llm": {
    "provider": "openai-compatible",
    "baseUrl": "https://api.swanchain.io/v1",
    "apiKey": "sk-swan-YOUR-API-KEY",
    "model": "zai-org/GLM-4.7-Flash"
  }
}

IronClaw (Rust)

Security-focused with TEE and encrypted vault. Built by NEAR AI.

export LLM_BACKEND=openai_compatible
export LLM_BASE_URL=https://api.swanchain.io/v1
export API_KEY=sk-swan-YOUR-API-KEY
export MODEL=zai-org/GLM-4.7-Flash

ironclaw

NemoClaw (TypeScript + Python)

NVIDIA's reference stack with kernel sandbox and privacy router.

Configure the inference provider in your NemoClaw deployment:

inference:
  provider: openai-compatible
  base_url: https://api.swanchain.io/v1
  api_key: sk-swan-YOUR-API-KEY
  model: zai-org/GLM-4.7-Flash

NemoClaw is in early preview (alpha). Configuration may change.

Moltworker (TypeScript, Cloudflare Workers)

Runs on Cloudflare's edge network (330+ cities).

{
  "models": {
    "providers": {
      "swan": {
        "type": "openai-compatible",
        "baseUrl": "https://api.swanchain.io/v1",
        "apiKey": "sk-swan-YOUR-API-KEY"
      }
    }
  }
}

Or set via Cloudflare AI Gateway:

AI_GATEWAY_BASE_URL=https://api.swanchain.io/v1

NullClaw

Ultra-minimal (678KB binary, 1MB RAM). For IoT and embedded devices.

nullclaw --provider "custom:https://api.swanchain.io/v1" \
         --api-key "sk-swan-YOUR-API-KEY" \
         --model "Qwen/Qwen2.5-7B-Instruct"

Or in config:

[provider]
url = "custom:https://api.swanchain.io/v1"
api_key = "sk-swan-YOUR-API-KEY"
model = "Qwen/Qwen2.5-7B-Instruct"

For resource-constrained devices, use smaller models like Qwen/Qwen2.5-7B-Instruct for faster responses.

Choosing the Right Tool

Scenario

Recommended Tool

Why

Quick local testing

ZeroClaw

3.4MB, boots in <10ms

Python project

Nanobot

pip install, native Python API

Multi-platform bot

OpenClaw

25+ messaging platforms

Edge / ARM device

PicoClaw

8MB Go binary, runs anywhere

IoT / embedded

NullClaw

678KB, runs on $5 hardware

Security-critical

IronClaw

TEE, encrypted vault, WASM sandbox

Multi-tenant SaaS

NanoClaw

Container-per-session isolation

Enterprise / NVIDIA

NemoClaw

Kernel sandbox, policy enforcement

Global edge deploy

Moltworker

Cloudflare Workers, 330+ cities

Playground (No API Key)

The public playground is a web UI and a bespoke endpoint (POST /v1/playground/chat), not an OpenAI-compatible base URL — agent tools cannot be pointed at it. Use it to try a model by hand, then sign up for a free sk-swan-* key; every tool on this page needs one.

Troubleshooting

Connection refused / timeout

Verify the base URL includes /v1: https://api.swanchain.io/v1
Check your API key starts with sk-swan-

Model not found

List available models: curl https://api.swanchain.io/v1/models -H "Authorization: Bearer sk-swan-YOUR-KEY"
Model IDs are case-sensitive (e.g., Qwen/Qwen2.5-7B-Instruct, not qwen2.5-7b)

Rate limited (429)

Default rate limit is 200 requests/min for LLM models
Check X-RateLimit-Remaining header in responses
If you are on pay-as-you-go, the per-category rate limits in the API reference apply; the $6/month Pro plan includes $24 of inference/month at list prices on free- and standard-tier models but has its own 50 requests/min limit, so it helps with cost, not burst rate

Streaming not working

Ensure your tool is configured for streaming (stream: true)
Swan Inference supports SSE streaming on /v1/chat/completions