12-factor-agents
12-factor-agents copied to clipboard
humanlayer
What are the principles we can use to build LLM-powered software that is actually good enough to put in the hands of production customers?
This introduction to the 12 key factors for building Agent applications is great, but I'd also like to see some examples of Agent development that puts these factors into practice....
I am sure you mean that the LLM figures out how to the connect the dots...
Great work summarizing the core principles — it’s a solid bridge between software deployment discipline and the emerging agent paradigm. One suggestion: we could explore evolving this into a “12-Factor...
## Summary Fixed an inconsistency in the example where the memo addressed "Jeff" instead of "Terri", even though the payment link was being created for Terri.
Just a minor addition to the "Compact Errors" section that error handling can often be removed from the agent's context window once a successful output is achieved. You demonstrate that...
This addition proposes to start adding evaluation tests (evals) to be part of the context of agent app. General concept: - https://platform.openai.com/docs/guides/evals-design#multi-agent-architectures - https://cookbook.openai.com/examples/evaluation/getting_started_with_openai_evals The main idea of the PR...
A template/boilerplate codebase for a production agent that implements the 12 factors.
Metadata
Owner
Metadata
What are the principles we can use to build LLM-powered software that is actually good enough to put in the hands of production customers?