Vivgrid

Vivgrid has a free plan and free credits, which makes it easy to try on a personal or small project before committing to anything bigger.

It is an AI agent platform that covers building, testing, deploying, observing and evaluating agents under one roof.

Inference runs on globally distributed infrastructure, aimed at low latency and reliable scale.

The observability side is where Vivgrid spends most of its effort.

You can see prompts, API calls and memory usage, and step through an agent's behavior to track down errors.

Deployed agents get monitoring for latency, cost and usage.

Automated evaluation, performance scoring and safety guardrails such as refusal rules and content filters are built in.

Multi-agent orchestration with stateful, context-aware memory is supported too.

Models from OpenAI, Anthropic Claude, X.ai and Google Gemini can all be used, so you are not tied to a single provider.