12-Factor Agents
Our verdict
A practical checklist for moving from demo to production: own your prompts, own your control flow, keep state explicit. Opinionated and useful.
- Who it's for
- Builders taking agents to production
- Leverage
- Fewer production surprises
- Effort
- Low
- Date added
What it does
12-Factor Agents is an open guide from HumanLayer setting out principles for building reliable LLM-powered software.
How builders use it
- Audit your agent against each factor.
- Pull control flow out of the prompt and into code.
- Design explicit human-approval steps.
What to expect
Agents that behave more like normal, debuggable software.
Related discoveries
- Multi-Step Agent SetupsUse case
Building effective agents: the patterns
Prompt chaining, routing, parallelisation, orchestrator-workers, evaluator-optimizer.
Effort: Low
- Multi-Step Agent SetupsTool
Model Context Protocol (MCP)
The open standard for connecting AI apps to tools and data.
Effort: Medium
- Multi-Step Agent SetupsTool
MCP reference servers
Official example MCP servers to use or learn from.
Effort: Low
- Multi-Step Agent SetupsTool
GitHub MCP Server
Give agents structured access to issues, PRs, code and Actions.
Effort: Low
Want more like this?
Practical discoveries, without the noise.