From the team behind Gatsby, Mastra is a framework for building AI-powered apps and agents with workflows, memory, streaming, evals, tracing, and Studio, an interactive UI for dev and testing. Start building: npm create mastra@latest
This is the 4th launch from Mastra. View more
Mastra Factory
Launching today
Mastra Factory is an open source, agent-powered software delivery environment. It combines persistent coding agents, repository workspaces, issue intake, planning, implementation, and pull request review in a web application you control.
Launch Team
GitHub
Claude by Anthropic
Langchain
crewAI
Mastra
Love it - @Mastra (YC W25) is building an agent-powered software factory, and it's open source.
What's a software factory? It's a system for agents to take software from issue into production.
@Spotify recently launched Xirp. @Mastra's is open source and configurable. You connect GitHub, you connect Linear and Slack, and you define rules for how work moves through phases. Agents triage issues, write and validate code, release changes, update docs, and monitor prod.
View source code on github.com
Read the full announcement on mastra.ai
Over to you:
@Mastra started as an agent framework. It's also a factory framework now.
S/O to @calcsam @smthomas3 @bookercodes and team for this new launch. Pumped to contribute to it.
S/O to the @Product Hunt community for all the love in the reviews.
@eldadafux from @Appwrite / @Imagine:
@flyakiet from @Superset:
@Mastra (YC W25) is building an open source agent-powered software delivery environment.
Star this repo: github.com/mastra-ai/mastra
And start building:
Mastra
@calcsam @smthomas3 @eldadafux @flyakiet @fmerian Thank you 🙇🏻
Mastra
Alex from the video here 👋🏻 If you’re not familiar, Mastra is a TypeScript framework for building AI agents. We used Mastra to build Mastra Factory, and now we’re using Factory to build Mastra. It’s already writing more than 25% of our pull requests!
Building agents is what we do at Mastra, and we’re invested in making this better. But most teams just want to build their product, not take on building and maintaining a factory too. That’s why we’re sharing ours.
With Mastra Factory, you can own the factory without having to build it.
Metabase
Full circle (in production!) Congrats on the launch 👏
Mastra
@ramiro__nd Ramiro! Thank you 🙏🏻
Mastra
@bookercodes #dogfooding
Mastra
@fmerian Literally!
Mastra
exactly - persistent coding agents, repository workspaces, issue intake, planning, implementation, and pull request review. all in a web application you control.
learn more: factory.mastra.ai
Mastra
@shivam_kushwaha16 Yeah, you can kick-off a long-running persistent agent, close your laptop, go to bed, then see how far it got on the goal in the morning.
@shivam_kushwaha16 persistent is the word i would push on here. most tools claiming it keep the transcript, and the transcript is the part that ages worst. the thing worth keeping is what got decided.
Mastra
@shivam_kushwaha16 @rabnoor_s Makese sense!
DiffSense
Whats the coolest thing you have seen ben built with this? # Proof is in the pudding
Mastra
@conduit_design So far, Mastra.
Mastra
@Mastra? the team built the factory first to automate their own issues. it's now writing 25-35% PRs and closing 50-60% issues.
view source code on github.com
So, I can use my Factory from Slack? I guess I'll have to cancel my Devin sub. Thanks for that!!
Mastra
@toolwanderer yes, the factory connects @GitHub, @Linear, and @Slack to configurable boards. it pulls in GH issues and pull requests, Linear issues, and Slack threads before an agent touches them and moves through phases.
learn more: factory.mastra.ai
Does it store context indefinitelyor does older data get trimmed automatically after some point?
@devinstone Funny enough what started this journey was us trying to make an agent that "never forgets". This led to Observational Memory https://mastra.ai/research/observational-memory. With that foundation we started going deeper into Coding Agents and now the factory. Also gave a talk on Factory here
one thing i would want to understand better is how memory behaves when the same agent is serving many different users.