Hecate FAQs

Question 1

What is Hecate and what problem does it solve?

Accepted Answer

Hecate is an open-source AI gateway and agent-task runtime designed to provide a unified control plane for accessing and managing both cloud and local AI models. It addresses the complexity of modern AI workloads by centralizing model access, cost governance, routing, caching, observability, and controlled agent execution.

Question 2

What kind of AI models and clients does Hecate support?

Accepted Answer

Hecate supports OpenAI-compatible and Anthropic-compatible clients, acting as a proxy. It seamlessly integrates with a wide range of model providers, including major cloud services like OpenAI and Anthropic, as well as local AI solutions like Ollama, LM Studio, LocalAI, and other llama.cpp-compatible servers.

Question 3

How does Hecate help manage AI costs and budgets?

Accepted Answer

Hecate offers comprehensive cost governance features. This includes spend controls, rate limits, an auditable pricebook, balance management, and usage tracking. It allows operators to gain full visibility and control over their AI expenditures, supporting both single-user and multi-tenant billing scenarios.

Question 4

What is the agent-task runtime feature?

Accepted Answer

The built-in agent-task runtime enables queued execution of AI agents, providing capabilities for approvals, retries, and resumable runs. It also includes sandboxed operations for controlled execution of shell commands, file system access, and Git operations, enhancing security and reliability for agent-driven automation.

Question 5

Does Hecate offer observability for AI workloads?

Accepted Answer

Yes, Hecate provides first-class OpenTelemetry integration for extensive runtime observability. It captures traces, metrics, and logs for every request, offering insights into route reports, failover details, costs, cache decisions, and overall system performance, crucial for debugging and optimization.

Hecate

Hecate

주요 기능

사용 사례

주요 기능

사용 사례