openclaw-nebius

name	openclaw-nebius
description	Nebius Token Factory provider plugin — adds 44+ open-source models (Qwen, DeepSeek, Llama, Gemma, GLM, FLUX, and more) to OpenClaw via a single OpenAI-compatible endpoint
version	1.3.1
homepage	https://github.com/opencolin/openclaw-nebius/tree/main/tokenfactory-plugin
metadata	{"openclaw":{"requires":{"env":["NEBIUS_API_KEY"]},"primaryEnv":"NEBIUS_API_KEY"}}
env	[{"name":"NEBIUS_API_KEY","required":true,"description":"Nebius Token Factory API key from studio.nebius.ai"}]

Nebius Token Factory Provider Plugin

Adds 44+ open-source models to OpenClaw via the Nebius Token Factory inference API.

Installation

Run this single command to install the plugin from ClawHub:

openclaw plugins install clawhub:tokenfactory

Then configure the API key and restart the gateway. Full steps below.

Step-by-Step Setup

Step 1: Install the plugin

openclaw plugins install clawhub:tokenfactory

If reinstalling, remove the old version first:

rm -rf ~/.openclaw/extensions/nebius
openclaw plugins install clawhub:tokenfactory

Step 2: Get an API key

Sign up or log in at https://studio.nebius.ai/ and create an API key.

Step 3: Set the API key

The gateway runs as a macOS LaunchAgent. Shell env vars from .zshrc are NOT visible to it. Use launchctl:

launchctl setenv NEBIUS_API_KEY "v1.YOUR_KEY_HERE"

Also create the auth profile for the agent. Add this file at ~/.openclaw/agents/main/agent/auth-profiles.json:

{
  "profiles": {
    "nebius:default": {
      "type": "api_key",
      "provider": "nebius",
      "key": "v1.YOUR_KEY_HERE"
    }
  },
  "lastGood": {
    "nebius": "nebius:default"
  }
}

If auth-profiles.json already exists, merge the nebius:default entry into the existing profiles object and add "nebius": "nebius:default" to the existing lastGood object.

Step 4: Add nebius to the plugins allowlist

openclaw config set plugins.allow '["nebius"]'

If other plugins are already in the allowlist, include them too (e.g. '["nebius", "openclaw-mem0"]').

Step 5: Restart the gateway

openclaw gateway restart

Step 6: Verify

openclaw plugins inspect nebius
openclaw models list --provider nebius

You should see Status: loaded and at least one model listed.

Step 7 (optional): Set a Nebius model as default

openclaw config set agents.defaults.model.primary "nebius/deepseek-ai/DeepSeek-V3.2"
openclaw gateway restart

Deploying OpenClaw on Nebius

To deploy OpenClaw as a serverless endpoint on Nebius AI Cloud, use the nebius skill. It covers the full deployment workflow — creating endpoints, configuring regions, setting up networking, and connecting via TUI or dashboard.

Required Credentials

Variable	Required	Description
NEBIUS_API_KEY	Yes	API key from studio.nebius.ai

No other credentials are needed. The plugin only communicates with https://api.tokenfactory.nebius.com/v1.

Available Models

All models use the nebius/ prefix. Examples:

• nebius/deepseek-ai/DeepSeek-V3.2 — DeepSeek V3.2 (163K context, chat)
• nebius/Qwen/Qwen3-235B-A22B-Thinking-2507 — Qwen3 235B (reasoning)
• nebius/Qwen/Qwen3-Coder-480B-A35B-Instruct — Qwen3 Coder 480B
• nebius/moonshot-ai/Kimi-K2.5 — Kimi K2.5 (262K context)
• nebius/meta-llama/Llama-3.3-70B-Instruct — Llama 3.3 70B
• nebius/google/Gemma-3-27b-it — Gemma 3 27B
• nebius/NousResearch/Hermes-4-405B — Hermes 4 405B (reasoning)
• nebius/openai/gpt-oss-120b — GPT-OSS 120B (reasoning)

38 chat/reasoning models and 2 image generation models (FLUX.1) are included.

See the full catalog in SETUP.md.

Troubleshooting

"plugin not found: nebius"

• Run openclaw plugins install clawhub:tokenfactory
• If reinstalling: rm -rf ~/.openclaw/extensions/nebius first

401 Unauthorized

• API key expired or wrong. Generate a fresh one at studio.nebius.ai
• Make sure you ran launchctl setenv NEBIUS_API_KEY "..." (not just export)
• Restart the gateway after changing the key

"Unknown model: nebius/..."

• Run openclaw gateway restart after installing
• Check openclaw plugins inspect nebius shows Status: loaded

Models show as "missing"

• The plugin catalog auth may not resolve. Add models to config directly: See SETUP.md section on config-based model registration.