跳到正文
Mastra Blog·· 6 天前AI 評分38

Mastra 推出 Custom Factory 計劃

Introducing Custom Factory Program

AI 導讀

Mastra 推出 Custom Factory 計劃,與客戶共同打造配合其程式碼庫及工程流程、在客戶技術堆疊上運行的軟件工廠。Mastra 自家的軟件工廠現已處理 70% 的 issues,推出後 PR 數量增至 6x。客戶可獲內置 Evals 和可觀測性、人手檢查點及安全治理,並擁有工廠及其產出。

正文

Today we're launching our Custom Factory program. Our forward-deployed engineers embed with your team and co-build a software factory for your codebase on Mastra, running in your stack.

We know what a factory can do because we run one. Mastra's software factory now handles 70% of our issues, and since we launched it, our PRs have grown 6x. Most of that work starts in Slack with Shipyard, our Slack agent. Anyone on the team can hand Shipyard a bug or a feature request, and the factory takes it from there, all the way to a pull request ready for review.

We built our factory for our organization, centered around Mastra as an open-source project. It's a great starting spot, and most engineering organizations have their own processes. So we're taking the factory as a starting spot, and tailoring it to your organization and your processes.

Mastra's factory board: issues are triaged, researched and moved from intake to review, with each one scored for impact and effort.

Most engineering teams are still adopting AI one tool at a time: a coding agent someone runs by hand, an issue investigation agent, a code review agent.

But a factory does the whole thing – it takes work from issue to merged PR. It triages and scopes the task, writes the code, runs the tests, checks its own output, and brings in an engineer where judgment matters. Each step in the pipeline is run by an agent – and it gets better every time it runs.

We've watched other teams build other kinds of factories, too. Our go-to-market systems run on a factory. Companies like Translucent have built factories for accounting.

If you're trying to get a factory in your org, we'd love to help.

What the program entails

You'll get a software factory designed around your codebase: your repos, conventions, CI, and review process.

This includes:

  • The underlying factory, and a Slack agent that lets you operate it
  • Forward-deployed engineers who embed in your repo and team chat, bringing patterns from the factory we run ourselves
  • Evals and observability built-in, so you can see what the factory is doing and prove it's working
  • Human checkpoints so engineers approve what ships, with security and governance built-in (auth, role-based access control, etc.)

Plus you own the factory and everything it produces, and your team can operate and extend it after launch.

To get there, we start with a whiteboarding session. We want to understand your codebase and organization before we build together. Then, we'll co-embed with your team and co-build it until it's shipping PRs in production.

Get started

Grab time to whiteboard with our team.

來源:Mastra Blog · mastra.ai