Charlie Labs has introduced Daemons, a new product designed to keep watch over software teams’ recurring maintenance tasks without needing repeated prompts from humans. The company says the system is meant to operate continuously across tools such as Slack, Linear, GitHub, and related workflows, handling routine follow-through while engineers focus on higher-value work.
Daemons are described as persistent AI processes that can be defined in Markdown files stored in a repository. Instead of assigning a narrow task with a clear start and finish, users specify a role, a purpose, the events it should monitor, the actions it can take, and the actions it must avoid. Charlie Labs says that approach is intended to make the behavior predictable and easier for teams to trust.
The product is aimed at recurring coordination and maintenance work that often gets delayed or overlooked. According to Charlie Labs, daemons can help organize pull requests, issue trackers, ownership, priorities, and team context so information stays aligned. They can also maintain code, dependencies, documentation, runbooks, and other operational material to reduce drift as a project changes over time.
The company’s examples include a PR helper daemon that watches pull requests as they are opened or updated, suggests improvements to descriptions, and flags missing reviewer context. Another sample daemon handles issue labeling in Linear by adding missing labels when new issues are created or when unlabeled items are found in a daily sweep. A separate bug-triage daemon watches for bugs, sets priority using Sentry impact data and context, assigns owners based on CODEOWNERS, and appends missing information to the issue body.
Charlie Labs says these daemons are deliberately constrained. They can be configured with deny rules that prevent them from taking actions such as merging pull requests, changing protected branches, deleting issues, or altering labels that are already in place. The company also points to rate limits and other limits sections in the daemon files to keep the work from overwhelming teammates or review processes.
The daemon definition format uses frontmatter for declarative fields such as name, purpose, watch rules, routines, denied actions, and schedules. Below that, additional Markdown can define policy, output format, escalation rules, and other operational guidance. Charlie Labs says the format is open and portable, allowing the same file to work across providers that support the specification.
The company is also framing Daemons as a complement to existing AI agents rather than a replacement. In its positioning, agents help build new things, while daemons keep existing systems maintained over time. Charlie Labs argues that the value comes from ongoing attention to repetitive work that teams otherwise need to keep rediscovering and reassigning.
The broader pitch is that once a daemon is added to a repository, the benefit extends across the team without a separate rollout or onboarding process. Charlie Labs says teams can define the role once and let the daemon accumulate context, improve its understanding of the codebase, and continue working in the background.
The launch reflects a growing push in software development tools toward autonomous maintenance systems that can operate with fewer prompts and less manual follow-up. Charlie Labs is betting that engineering teams will see value in AI that handles the unglamorous but necessary work of keeping projects tidy, current, and review-ready.