Core Architectural Capabilities
Engineered for Speed, Precision & Complete Cost Control
Everything you need to deploy, meter, and scale multi-model AI workloads with enterprise security.
Dual Protocol Wire Proxy
Point OpenAI client SDKs directly to /v1/chat/completions or native Ollama tooling (Zed IDE, Cursor, OpenCode) to /api/generate with zero proxy friction.
- OpenAI v1 Compatible REST
- Native Ollama JSON & Streaming
- Single-Model Auto-Route (262k Context)
GPU Wattage & kWh Metering
Real hardware power load monitoring with millisecond precision. Assign custom £/kWh rates to clusters and pay strictly for physical compute energy consumed.
- Live Wattage & Latency Auditing
- Prepaid Wallet & Monthly Invoicing
- Fractional-Penny Deduction Ledger
Reasoning & Live Web Tools
Toggle thinking depth budgets (Low, Medium, High), upload up to 512MB documents and images via dropzone, and execute multi-round live website browsing.
- Collapsible Thought Process Accordion
- Full-Screen 512MB File Dropzone
- Autonomous Web Page Fetching Tool
Multi-Model Ecosystem
Supported Upstream Providers & Failover Hierarchy
Individual profile-level failover hierarchies with dedicated model priorities and granular model disablement.
Enterprise Invoicing & Multi-Processor Payments
Manage team spending limits, schedule restrictions, custom profit markups, and billing agreements. Support instant balance top-ups and online invoice settlements via Stripe, PayPal, SagePay/Opayo, or BACS direct transfers.
from openai import OpenAI
client = OpenAI(
base_url="https://api.ai.uk1.handywebsites.co.uk/v1",
api_key="sk_live_..."
)
response = client.chat.completions.create(
model="auto",
messages=[{"role": "user", "content": "Analyze this dataset"}]
)
print(response.choices[0].message.content)