OpenClaw API Key Not Working — How to Fix It 2026

API key errors stop your agent cold. Here's how to systematically diagnose and fix every common API key issue in OpenClaw.

Step 1: Test the API Key Directly

# Test Anthropic key:
curl -s -H "x-api-key: $ANTHROPIC_API_KEY" \
  -H "anthropic-version: 2023-06-01" \
  -H "Content-Type: application/json" \
  https://api.anthropic.com/v1/messages \
  -d '{"model":"claude-haiku-3-5","max_tokens":10,"messages":[{"role":"user","content":"hi"}]}'
# Expected: a short response. Not a 401/403.

# Test OpenAI key:
curl -s https://api.openai.com/v1/models \
  -H "Authorization: Bearer $OPENAI_API_KEY"
# Expected: list of models

# Test Google Gemini key:
curl -s "https://generativelanguage.googleapis.com/v1beta/models?key=$GOOGLE_API_KEY"
# Expected: list of models

Step 2: Check the .env File

# Verify .env exists and key is set:
cat ~/.openclaw/.env | grep -i "api_key\|token"

# Common mistakes:
# Wrong variable name:
ANTHROPIC_KEY=sk-ant-...          # WRONG
ANTHROPIC_API_KEY=sk-ant-...      # CORRECT

# Extra spaces around = sign:
ANTHROPIC_API_KEY = sk-ant-...   # WRONG
ANTHROPIC_API_KEY=sk-ant-...     # CORRECT

# Quotes around key (usually not needed):
ANTHROPIC_API_KEY="sk-ant-..."   # May cause issues
ANTHROPIC_API_KEY=sk-ant-...     # Correct

Step 3: Verify OpenClaw Config Matches Provider

# Check current config:
openclaw config list | grep llm

# Make sure provider and model match:
openclaw config set llm.provider anthropic
openclaw config set llm.model claude-haiku-3-5

# For OpenAI:
openclaw config set llm.provider openai
openclaw config set llm.model gpt-4o-mini

# For Google:
openclaw config set llm.provider google
openclaw config set llm.model gemini-2.0-flash

# Restart after config change:
openclaw gateway restart

Step 4: Check Key Permissions and Scopes

Some API keys have restricted permissions:

• Anthropic: No scope restrictions — key either works or doesn't
• OpenAI: Project keys can have restricted model access. Check at platform.openai.com → API Keys → permissions
• Google: Ensure the Generative AI API is enabled in your Google Cloud project and the key has access to it

# OpenAI: check if model is accessible with your key:
curl -s https://api.openai.com/v1/models \
  -H "Authorization: Bearer $OPENAI_API_KEY" | jq '.data[].id' | grep gpt-4

Step 5: Check for Key Expiration

# Anthropic and OpenAI keys don't expire by default
# But if you set an expiry date when creating:
# Anthropic: console.anthropic.com → API Keys
# OpenAI: platform.openai.com → API Keys
# Check the expiry column in the dashboard

Step 6: Check Billing / Usage Limits

# Error 429 with "billing" in message = you've hit spending limit or ran out of credits

# Anthropic: Add credits at console.anthropic.com → Plans & Billing
# OpenAI: Add credits at platform.openai.com → Usage → Add payment method
# Google: Check quota at console.cloud.google.com

Step 7: Gateway Restart After Fix

openclaw gateway restart
openclaw gateway status
# Should show: LLM: ok

Get the complete setup guide in The OpenClaw Playbook — everything you need to master OpenClaw for $9.99.