AGENT PLAYGROUND
Test your self-evolving agent live.
Paste your API key, customize your agent, and watch it learn from every message.
Enter your API key
Your key runs your agent. It stays in your browser only — never sent to our servers.
Groq
Free, ultra-fast inference
FREE
OpenAI
GPT-4o, GPT-4 Turbo
POWERFUL
OpenRouter
100+ models, one key
MULTI-MODEL
Claude (Anthropic)
Claude 3.5, Claude 3 Opus
CLAUDE
Custom
Ollama · LM Studio · Together AI · Fireworks · Mistral · Perplexity · Anyscale · Deepinfra
BYOB
Get your free key at console.groq.com — free tier: 14,400 req/day
🔒 HOW YOUR KEY IS USED
Your key never leaves your browser. It calls the API directly — no proxy, no server in between. You can verify this by opening DevTools → Network tab and checking the request goes straight to the provider.
// This is exactly what happens in your browser:
fetch('https://api.groq.com/v1/chat/completions', {
headers: { 'Authorization': 'Bearer ' + YOUR_KEY }
})
// No evoclaw.tech server involved ✓
fetch('https://api.groq.com/v1/chat/completions', {
headers: { 'Authorization': 'Bearer ' + YOUR_KEY }
})
// No evoclaw.tech server involved ✓
Configure your agent
Customize your agent's name, model, and personality.
Llama 3.3 70B
Best quality, longer responses
SMART
Llama 3.1 8B
Fast, great for quick replies
FAST
Mixtral 8x7B
Good for technical topics
TECHNICAL
Gemma 2 9B
Google's efficient model
EFFICIENT
My EvoClaw Agent
llama-3.3-70b-versatile
Your agent is ready.
Ask it anything — it learns from every message.
AGENT CONFIG
Model
llama-3.3-70b
Personality
assistant
Memory
● ACTIVE
Evolution
● ON
PRM SCORE
—
PROCESS REWARD MODEL
SESSION STATS
0
MESSAGES
0
SKILLS
SKILLS INJECTED
Waiting for first message...
🧠 JUST LEARNED
Nothing yet...