schemaVersion: v1alpha1name: web-server-factorydescription: Autonomous software-development factory for a web application.
alias: web-factoryThe alias is the Foreman’s @-mention handle in Slack and Linear, so teammates can route work to the Factory by name. It can differ from the Factory name.repositories: - owner: acme name: web-server - owner: acme name: web-clientRepositories define where Factory agents can work. Each run checks out the relevant code on its execution host so agents can inspect and change it.secrets: - DEPLOY_API_TOKENSecrets give agents protected credentials for the tools and services they need. Values are encrypted at rest and injected as environment variables only during runs.mcpServers: sentry: warpId: YOUR_MCP_SERVER_IDMCP servers connect every agent to shared tools such as Sentry or internal services. Each named connection references a centrally managed server.cloudProviders: aws: roleArn: arn:aws:iam::123456789012:role/factory-agentCloud providers securely connect Factory agents to cloud accounts and resources for infrastructure work. Federated access lets runs assume an AWS role or use a GCP identity without embedding long-lived cloud keys in the definition.integrations: - type: slack - type: linearIntegrations connect the Factory to work sources such as Slack and to Linear or Jira for task tracking, so the Foreman can coordinate work where your team already works.agentDefaults: model: auto runner: standardAgent defaults establish the model or harness, runner, and environment that roles inherit, while specialized agents can override those choices when needed.Measurement > Vibes
Warp Factories is open infrastructure for teams to build measurable, self-improving software factories.
▩queued■working▨waiting□shipped
224 tasks · 0 active · 1,422 PRs shipped
Trusted by 800k+ devs at
Your software factory, defined in code.
Factories-as-code allows for a versioned and measured approach to optimizing costs and performance of your factories.
Benchmark. Optimize. Repeat.
Benchmark on your own work
Replay historical tasks from your factories to measure performance against your team’s real workflows.
Measure what performs best
Compare models, harnesses, and configurations across quality, correctness, efficiency, and custom scorers.
Put the results to work
Update your factory code to route work to the best-performing setup.
Beyond CI/CD to automating the whole SDLC
defined in code, built for scale, and easy to deploy.
How work moves through a factory
issues land from the channels your team already uses, the foreman moves them down the line, and the line stops wherever you decide.
[ fig. t10 · issue flow · the pipeline, with the foreman ]option 1b · foreman
slackcheckout button broken on mobile safari
the foreman carries this issue down the line — watch for the f badge
01 triage
optional
cause, scope, a tracked issue
02 spec
optional
the plan for how the change should land
│[review the spec]
needs spec approval
03 implement
writes the code and tests, opens a pr
│[review the code]
needs code review
04 verify
adversarial review, evals, and checks
│[review the product]
needs qa
05 ship
the approved change lands
06 monitor
watches prod, files what it finds
the foreman moves the work · it stops only where you decide · merge is always yours
Cutting our cost per PR from $80 to $30
How we used Warp Factories Benchmarks to test models on our own engineering tasks and optimize for cost without sacrificing quality.
read case study
Open at every layer
Layer 01
Any agent
any MCP-capable coding agent.
- warp
- claude code
- codex
- cursor
Layer 02
Any model
frontier or open-weight, chosen per pipeline stage.
- frontier
- open-weight
Layer 03
Your compute or ours
warp's cloud, or self-hosted in your own VPC.
- warp cloud
- self-hosted
Layer 04
Data lives where you want
you own and store what your factory produces.
- pluggable
- zero-retention
- your VPC
Baked-in measurement and self-improvement
evals, benchmarks, and self-improvement loops drive measurable gains.
EVALpass
Evals on your own work
96%
8/13 07:03fail0.4
8/13 07:02pass0.8
8/13 07:02pass0.8
BENCHbest
Benchmarks across models
1.00x
Claude Fable 5pass1.00x
GPT-5.6 Solpass0.94x
Gemini 3.6 Flashfail0.89x
LOOPauto-fix
Self-improvement loops
+3.7
memory updatedpass+0.2
prompts tunedpass+3.2%
regression caughtfail-0.4
"warp factories drove our cost per agent pr down by 30%."
— vp engineering, series c infrastructure company
Control your coding agent chaos
continuous improvement, better governance and security, by default.
increase agent roi · cost per pr
$—
computeplatforminference
[ fig. t6 · agents · live roster ]3 agents
ready for taskGPT 5.2 CODEXCODEXrnow
drafting the specOPUSCLAUDE CODEs14m
routed to specGPT 5.2WARPt37m
Status advances on a shared clock; no agent is ever idle.
Agents prove their work with computer use
every factory agent captures a screenshot or video so you can verify its work before shipping a PR.
The factory infrastructure you'd build yourself
available now, so your eng time goes into your product, not the plumbing. tick a box to clear the desk (drag the pile around on desktop).
TO_DO
Slack and Teams event webhooks
TO_DO
Jira and Linear issue webhooks
TO_DO
GitHub and GitLab push/PR webhooks
TO_DO
webhook signature verification
TO_DO
cron scheduler for triggered runs
TO_DO
event-to-task deduplication
TO_DO
per-run sandbox provisioning
TO_DO
Docker- or Kubernetes-backed compute
TO_DO
environment image build and caching
TO_DO
cold-start latency for new runs
TO_DO
sandbox teardown and cleanup
TO_DO
multi-harness process management
TO_DO
Claude Code, Codex, Cursor, and Warp support
TO_DO
model routing per pipeline stage
TO_DO
MCP server connection management
TO_DO
CLI invocation inside the sandbox
TO_DO
source-forge integration: GitHub, GitLab
TO_DO
non-human identity per agent
TO_DO
credential and token handling
TO_DO
fine-grained permissions per agent
TO_DO
secrets storage and rotation
TO_DO
quotas and concurrency limits
TO_DO
network egress policy enforcement
TO_DO
run lifecycle state machine
TO_DO
durable execution across long-running steps
TO_DO
retries and idempotency
TO_DO
multi-agent handoff between stages
TO_DO
workflow versioning
TO_DO
thread and run continuity
TO_DO
real-time log and event streaming
TO_DO
run history and audit trail storage
TO_DO
audit trail per tool call
TO_DO
live session attach for steering
TO_DO
failure detection and alert routing
TO_DO
approval-gate enforcement before merge
TO_DO
checkpoint pause/resume state
TO_DO
evals and scoring pipeline
TO_DO
artifact and output storage
TO_DO
benchmark regression detection
TO_DO
…and everything you haven’t hit yet
Frequently asked questions
a factory is a fleet of agents wired to your SDLC — triggered by an issue, a slack message, or a schedule, and coordinated by warp factories from triage through review to a mergeable PR.
as code, in a factory.yaml + supporting agent files. define triggers, agents, models, and approval gates — check it into your repo like any other config.
no. bring your own model or harness (e.g. claude code or codex) — warp factories works with whatever your team prefers and helps identify the most cost-effective configuration over time.
no. factories is a separate product — nobody at your company has to use warp terminal. work comes in through tools you already use, like slack, linear, jira, github, and gitlab. if you do use warp terminal, it has built-in factory integrations (native mcp) so you can iterate locally.
most orgs start around 20-30% of PRs fully automated, starting with simple tasks. over time this goes up as models improve and your factory self-improves.
wherever you choose — warp's cloud, or fully self-hosted inside your own VPC, under your existing retention and compliance rules.
humans put work into the factory via your existing tools like slack, teams, linear, jira, github and gitlab. at any time you can iterate on that work locally and the factory agents loop you in proactively when they need help.
yes — most teams start with a single low-risk workflow, like dependency bumps or flaky test triage, before expanding coverage.
usage-based, priced per agent run.