We started LoopLlama in 2024 with a narrow bet: that teams wanted to ship agentic features without becoming experts in agent orchestration. After a year of building alongside early users, that API is now generally available to everyone.
What's in GA#
At its core, GA does what we set out to do. You define a workflow — a reusable configuration with a crew of agents, each with a role and a system prompt — and trigger runs over a simple REST API. We handle the sequencing, the model calls, persistence, and usage metering. Every run is recorded step by step so you can see exactly what happened.
- Composable crews of agents with per-agent roles and prompts.
- Runs over a single REST endpoint, with job IDs and usage metering.
- Per-step records for every run, from the first day.
- Usage-based pricing — you pay for the steps your agents run, with a free tier to start.
Where we go from here#
GA is a starting line. The roadmap is full — connectors to real systems, human-in-the-loop, model routing, deeper observability — but the foundation is here today, and it's stable enough to build on. Get an API key and run your first crew; we're just getting started.