Automate boring code maintenance.
MassFactor is an AI code tool, a bit like GitHub Copilot, but it can work on thousands of files by using AI agents.
Instead of you chatting with an AI to change one file, the AI agents chat with each other to change a thousand files.
this is how massfactor helps you
Turn-Key Automations
Perform common maintenance tasks
on a massive scale, fully automatically.
MassFactor comes with ready to use, task-specific automations. These define which AI agents to deploy and how they collaborate.
Change / Refactor
This is the most generic automation. You give a natural language instruction and MassFactor performs it on all (or selected) files, optionally running tests afterward.
Sync changes between languages
If you're building a native iOS+Android app or a library for a dozen languages, you can add a feature in one language, and MassFactor syncs it to the others.
Modernize to a newer language version
Choose the language features you want to start using, and MassFactor handles the rest (e.g., change an entire JavaScript application to use async/await).
Translate to a new language #wip
AIs are good at translating code, but full codebase translation is hard. This automation is not (yet!) 100%, but can be useful as a starting point. Work in Progress.
Upgrade to a newer library version
After you upgrade a 3rd party library to the latest version, and things explode, MassFactor can fix it via reading "breaking changes" notes or your guidance.
Migrate to a different library
If the development of a 3rd party library you use has been discontinued, MassFactor can migrate your codebase to use a replacement library.
Review code & implement recommendations
The first agent does a code review, on a commit/PR or one file at a time on old code. The second agent implements the recommendations that you approve.
Analyze inconsistencies & unify patterns
You describe what to look for, such as an inconsistent design or naming pattern. The first agent analyses your codebase. The second unifies the patterns.
Generate tests
Choose classes/files to cover or a project-wide target coverage percent, and MassFactor generates the missing test, using the same style as in existing tests.
Generate docs #wip
In theory, this should work similarly to generating new tests, but we are struggling to get any LLM to produce insightful docs beyond the obvious. Work in progress.
Custom Automations
Execute complex custom logic, beyond maintenance tasks.Execute complex custom logic.
Design custom automations by reusing and wiring together our AI agents. As a simple example, the first agent can use a cloud LLM provider like OpenAI to modify a file, and then the second agent tests that it works.