Skip to main content
Common questions about Oqoqo. If you do not see yours, email support@oqoqo.ai. For setup failures, see Troubleshooting.
Evals and custom benchmarks for real work. See whether agents can use your product, compare agents and models on the same tasks, and find where they fail. See What is Oqoqo.
Yes. Organizations start on the free plan, which resets runs each month. You spend one run every time an agent tries a task under a treatment. Unused plan runs do not roll over. No credit card is required to start. See Pricing and Billing.
Subscribe for more runs each month from Organization settings → Usage & billing. Or top up anytime for runs that never expire. Larger top-ups cost less per run. Contact us for higher volume. See Pricing.
Runs cover Oqoqo platform fees for durable experiment workflows and cloud infrastructure. Model inference is separate. Use your own API keys or subscriptions.
Yes. Every plan includes unlimited members and projects, unlimited experiments and assets, and full web, CLI, and MCP access.
Files and a machine attach to a task. Skills, MCP servers, CLIs, and SDKs attach to a treatment. Agents are chosen at launch. See Core concepts.
Claude Code, Codex, Cursor, GitHub Copilot, OpenCode, Grok Build, Qwen Code, OpenClaw, Pi, Hermes, and Antigravity today. Custom agents are coming soon. See Supported agents.
Yes. Connect Anthropic (Claude Subscription) under Profile → Integrations, then enable Claude Code under Manage → Agents. You can also use an organization Anthropic API key. See Supported agents.
Yes. Connect Anthropic (Claude Subscription) under Profile → Integrations, then assign a Claude model under Oqoqo AI features. Ask Oqo cannot use this Integration yet. See Oqoqo AI features and Ask Oqo.
Connect a Cursor API key under Organization settings → Model providers, then enable Cursor under Manage → Agents. Cursor needs a paid Cursor plan. See Supported agents.
Connect an xAI API key under Organization settings → Model providers, then enable Grok Build under Manage → Agents. Grok Build uses xAI models only. See Supported agents.
Connect a Google Gemini API key under Organization settings → Model providers, then enable Antigravity under Manage → Agents. Antigravity uses Gemini models only. See Supported agents.
Connect Alibaba Model Studio or another compatible organization provider under Organization settings → Model providers, then enable Qwen Code under Manage → Agents. Use a pay-as-you-go Alibaba Model Studio key. Coding Plan and Token Plan endpoints cannot run experiments. Qwen Code does not use Cursor keys, GitHub Copilot, ChatGPT Subscription, Claude Subscription, or Nous Portal. See Supported agents.
Organization providers include OpenAI, Anthropic, Cursor, xAI, Google Gemini API, Alibaba Model Studio, DeepSeek API, Kimi API, Meta Model API, MiniMax API, OpenRouter, Vercel AI Gateway, Cloudflare AI Gateway, GMI Cloud, OpenCode Zen, OpenCode Go, and OpenAI- or Anthropic-compatible gateways. Personal Integrations include GitHub Copilot, OpenAI (ChatGPT Subscription), Anthropic (Claude Subscription), and Nous Portal. See Model providers.
Ask Oqo is the in-app assistant for your project. Assign a compatible model under Oqoqo AI features. You can use OpenAI (ChatGPT Subscription), Nous Portal, or another compatible provider. See Ask Oqo.
Use Ask Oqo in the dashboard, install the recommended plugin for your coding agent, or install the CLI. Use the dashboard for Matrix, Frictions, run detail, and secrets.
Max Concurrency limits how many agent sessions can run at once in a launch. It does not control Evals or Insights. See Launch options.
Evals score each run against your rubric. Insights analyze traces and output to find frictions. They are independent. See Evals and Insights.
Trying once is not an experiment. Oqoqo repeats the same tasks across agents, models, and treatments, then reports pass rates, lift, tokens, and frictions.
An ad hoc local run just executes the agent. Oqoqo runs the experiment around it: tasks with pinned machines and files, treatments with skills and tools, controlled comparisons, full trajectories, and pass or fail per requirement.
Skills, MCP servers, CLIs, SDKs, APIs, docs, files, and the workflows that use them. Attach interfaces to treatments and compare against the raw agent. See Assets.
Yes. Write a rubric of plain-language requirements. Evals score each run with an automated judge. See Tasks and rubrics.
Yes. Open any run and select Traces. Use Output for the final message and recorded workspace changes. Use Evals for per-requirement verdicts. See Traces and output.
Yes. From a run, export the transcript. Matrix and run detail show pass rate, lift, metrics, and evals. The CLI and MCP can also read runs, trajectories, Evals, and Insights.
Yes. Home includes templates such as Getting started, Stripe interface evals, Latest model shootout, and Skill lift. See Experiment templates.
Yes. On a finished experiment, choose Clone experiment to open a new draft with the same tasks, treatments, and settings. Review before you launch. See Launch options.
Yes. While an experiment is running, choose Cancel experiment. Unfinished runs stop. Completed runs stay available.
Yes. Select 2 or 3 runs on the Runs list and choose Compare. See Reading results.
Yes, from public github.com URLs. Use From GitHub in the import dialog, then Check for updates later. Private-repo import is not available. See Assets.
Plugins are coming soon under Manage → Assets → Plugins. Skills, MCP servers, CLIs, and SDKs are available today. See Assets.
Schedules are coming soon under Automate → Schedules. Today you can launch on demand from the dashboard, CLI, or MCP.
There is no first-class CI pipeline gate yet. You can launch and watch experiments from the CLI or MCP in your own automation.
Use the header project switcher for projects. Use Organization settings for shared org setup. Use Profile for Integrations, security, and email preferences. See the Dashboard map.