The platform for building
RL environments
Turn your production system into an environment.
Evaluate and train agents on your own software, web apps, or chat interfaces. Start from one of our templates and deploy in less than 30 minutes.
uv tool install hud-python
hud init
claude-sonnet-4-5 successfully performing a financial analyst task. Read the SheetBench case study →
One API for testing any model.
Stop juggling API keys. Point any OpenAI-compatible client at inference.hud.ai and use Claude, GPT, Gemini, or Grok. Every call is traced on hud.ai.
from openai import AsyncOpenAI
import os
client = AsyncOpenAI(
base_url="https://inference.hud.ai",
api_key=os.environ["HUD_API_KEY"]
)
response = await client.chat.completions.create(
model="model-name",
messages=[{"role": "user", "content": "Hello!"}]
)
print(response.choices[0].message.content)Run evaluations and training at scale.
Our infrastructure handles 1000s of concurrent environments with sub-second latency. Run full benchmark suites in minutes, not hours.
hud eval SheetBench-50 claude --remote
OSWorld-Verified
11050200
100×J×3=X traces
Pricing
SDK
Free
- ✓Turn any software into agent tools
- ✓Define scenarios for evaluation
- ✓Compatible with any agent framework
Cloud
$0.25+/environment hour
- ✓100+ parallel environment instances
- ✓Live telemetry and debugging
- ✓Detailed trace analysis
Start with $10 in free credits!
Start evaluatingEnterprise
Custom
- ✓Train agents on your environments
- ✓SOC 2 compliant infrastructure
- ✓Volume pricing and dedicated support
Are you a student or researcher? Get $100 in free credits with a .edu email. Making an academic eval? Apply for a grant.