Skip to main content
Oqoqo is the experimentation platform for the agentic era. Define a workflow once, run it across agents, models, and treatments in isolated production-like sandboxes, and measure exactly what changes — from token spend to friction points to eval outcomes. Whether you’re building the interfaces agents use or the workflows they run, Oqoqo gives you repeatable, comparable results at scale.

Introduction

Learn what Oqoqo is, who it’s for, and how the core experiment loop works.

Quickstart

Set up and launch your first agent experiment in minutes.

Core Concepts

Understand experiments, trials, treatments, traces, rubrics, and more.

Use Cases

Explore what teams are testing and measuring with Oqoqo.

​
Get started in three steps

You can go from zero to a running experiment in a few minutes. Here’s the shape of the journey.
1

Join the waitlist and get access

Go to oqoqo.ai/waitlist and request access. You get up to 100 runs free — no credit card required to start.
2

Define your first experiment

Write a natural-language task, attach a rubric, select your agents and treatments, and configure your library and environment. Oqoqo turns that configuration into a repeatable experiment you can rerun any time.
3

Launch trials and measure results

Set the number of parallel trials and click Launch. Oqoqo runs each trial in a clean, isolated sandbox, captures the full trajectory, and surfaces metrics, diffs, frictions, and eval outcomes for every run.

​
Explore use cases

Agent Evals

Turn real workflows into repeatable evals across agents, models, and treatments.

CI/CD Gating

Block pull requests that break agent workflows, not just unit tests.

Interface Testing

Improve the token, latency, and path efficiency of your MCP servers, skills, CLIs, SDKs, and APIs.

Drift Detection

Rerun the same workflows on a schedule as models, dependencies, and your product change.