What it does
Warp presents itself as infrastructure for running coding agents. The product list names three parts: Warp Terminal, described as a modern terminal for agentic coding; Warp Agent CLI, a coding agent that works in any terminal; and Warp Factories, for running fleets of coding agents across the software development lifecycle. A Factory is defined in code. A factory.yaml file connects repositories, secrets, MCP servers, cloud providers such as AWS, and integrations like Slack and Linear or Jira, and sets the model and runner that agents inherit. The site adds that a Factory can be reached through an API, CLI, SDK and MCP, and that work can be watched or steered from web, mobile, terminal or IDE. The site also describes a quality loop in which runs are scored with evals, models are compared on benchmarks, and observer agents open pull requests against the factory configuration. Listed use cases include first-pass code review, bug investigation, refactors and migrations, and incident response. The homepage links to docs, a pricing page and a GitHub repository, but states no prices itself.
Capabilities
- Warp Terminal for working with coding agents
- Warp Agent CLI that runs in any terminal
- Factories defined as code with agents, runners, skills and scorers
- Evals and benchmarks to compare models on your own work
- Slack, Linear and Jira integrations plus API, CLI, SDK and MCP access
When to use it
- You want a terminal built around agentic coding
- You want agents to review pull requests or investigate bugs across your repositories
- You want to measure agent quality and cost across different models
