Do I need to use Warp Terminal to use Factories?

No. Factories is a separate product. Work can be initiated from tools you already use (Slack, Jira, GitHub), though the terminal has built-in integrations for local iteration.

Can I use my own AI models or coding agents?

Yes. Warp supports frontier and open-weight models, as well as other MCP-capable harnesses like Claude Code or Codex.

Where does our code and factory data live?

You choose. It can live in Warp's cloud, or be fully self-hosted in your own VPC under your existing data retention and compliance rules.

How is a factory defined and configured?

As code. A factory is defined in a `factory.yaml` file and supporting agent markdown files, which specify triggers, agent roles, models, and approval gates. This is checked into your repo.

Directory/Coding/AI Assistants/Warp

Warp

Warp is an AI-native development platform centered on a modern terminal application built for agentic workflows. It combines a high-performance terminal with integrated AI capabilities, including an agent that can execute commands, write code, and perform complex development tasks. The platform also offers Warp Factories, an infrastructure for building and running fleets of coding agents across your software development lifecycle (SDLC). Key capabilities include an AI agent that understands context from your codebase, multi-agent orchestration via Factories, deep integrations with developer tools like GitHub and Slack, and enterprise-grade security and governance features. It's designed for developers, engineering teams, and enterprises looking to leverage AI to automate repetitive tasks, improve code quality, and accelerate their development workflows.

Coding

Visit website

Added on

Sep 21, 2026

Modern TerminalAI AgentCloud FactoriesBYO ModelsCredit SystemEnterprise SSOSelf-Hosted OptionCLI & SDK

Product information

Everything worth knowing about Warp

The complete picture — from everyday features to the technical detail builders and IT folks dig for.

Core features

What it actually does

Modern Terminal with AI Integration

A terminal built for AI-assisted development, featuring blocks, modern text editing, and native agent integration.

Warp Agent & Agent CLI

Run a powerful coding agent directly in the terminal or in the cloud, with access to frontier and open-source models.

Warp Factories (Early Access)

Define and run fleets of coding agents across your SDLC with infrastructure-as-code (factory.yaml).

Warp Drive & Collaboration

Unlimited objects, notebooks, and workflows for team knowledge management and shared context.

Technical capabilities

Credit-Based Usage System

Pay-as-you-go model using credits for agent inference, with volume-based discounts and team spend caps.

Multi-Harness & Model Support

Bring your own models (OpenAI, Anthropic, etc.) and coding harnesses (Claude Code, Codex, etc.) for flexibility.

Must watch videos

You should have an agent that grades your agents

6:22

Tutorial6:22

You should have an agent that grades your agents

Try Warp Factories: warp.dev/factories/request-access Let's measure the quality of coding agents (Claude Code, Codex, etc) with scoring agents. These use LLM-as-a-judge to measure output quality along metrics you care about, like code quality, efficiency, verbosity, and task compliance. These feed self-improvement and benchmarking workflows to improve your software factory automatically.

The secret to cheaper, smarter coding agents

3:44

Tutorial3:44

The secret to cheaper, smarter coding agents

With Warp Factories Benchmarking it’s now possible to build custom coding agent benchmarks (Claude, Codex, Grok, Kimi, etc) on your team’s real data and workflows with a few clicks. This is a deep dive of Factory Benchmarks: how to read the report, pick the right models, and build custom routers to reduce cost-per-PR.

How to build a model bench on your own coding tasks

6:06

Tutorial6:06

How to build a model bench on your own coding tasks

Factory Benchmarks are the model bench generated from your own agent coding tasks. They help you measure, test and improve coding agents by replaying past agent runs. We used these benchmarks to cut our cost-per-PR by 63% Get early access: warp.dev/factories/request-access

Who uses it

Real use cases, no hype

Automated Code Review

Run first-pass reviews on every PR using configured factory agents.

Bug Investigation & Routing

Automatically reproduce bugs, investigate root causes, and route fixes.

Large-Scale Refactoring

Scope, plan, and execute code migrations across a codebase.

Incident Response

Investigate production alerts, gather context, and summarize next steps.

Dependency Management

Automate routine tasks like dependency bumps and flaky test triage.

AI Model Benchmarking

Test different models on your own engineering tasks to optimize cost and quality.

What's great

Modern terminal with native AI agent integration.
Factories as code for full SDLC automation.
Bring your own model or harness flexibility.

Technical strengths

Built-in evals, benchmarks, and self-improvement loops.
Multi-harness support (Claude Code, Codex, etc.) via MCP.

Where it falls short

Pricing tiers can get expensive for high-volume agent usage.
Factories product is in Early Access, not generally available.
Enterprise features like self-hosted agents require custom pricing.

Technical limitations

Agent memory and cross-harness context are still in Research Preview.

For technical folks

The deep-dive specs

FAQ

Includes technical Q&A

A factory is a fleet of agents wired to your SDLC, triggered by events (like a PR or Slack message) and coordinated from triage through review to a mergeable PR.