Show HN: Adrafinil – keep a lid-closed Mac awake only while agents work

Description and use

**Brief Summary (2‑3 sentences)**
*Adrafinil* is an open‑source macOS utility that prevents a lid‑closed Mac from sleeping only while specified “agents” (e.g., background scripts, AI workers, or CI jobs) are running. It lets developers prototype AI‑enhanced workflows—such as Retrieval‑Augmented Generation (RAG) pipelines or autonomous agents—without keeping the whole machine awake all the time.

**Value Proposition**

• **Targeted wake‑locks**: Saves power and reduces wear by keeping the Mac awake only when needed, a common requirement for long‑running AI model inference or data‑fetching agents.
• **Zero‑model overhead**: The tool does not ship its own ML stack; it simply provides the OS‑level guard that lets you plug in any existing model or agent framework (LangChain, LlamaIndex, etc.) and focus on the AI logic.
• **Fast prototyping**: With a few configuration lines you can spin up a RAG service or an autonomous agent on a MacBook, test it locally, and iterate quickly before moving to cloud or containerized environments.

**Practical Adoption Path**

| Step | Action | Reason |
|------|--------|--------|
| 1️⃣ | **Clone & inspect** the repository; verify the license (MIT/Apache) and read the README. | Ensures legal compliance and that the code matches your security posture. |
| 2️⃣ | **Run the test script** (make test or ./run.sh) on a non‑critical Mac to confirm the lid‑closed wake‑lock works as advertised. | Quick sanity check before integrating into any pipeline. |
| 3️⃣ | **Define “agents”**: create a small wrapper script (e.g., agent.sh) that launches your AI process and exits when it finishes. | The utility monitors the wrapper’s PID to decide when to allow sleep. |
| 4️⃣ | **Configure** via the provided config.yaml (or CLI flags) to point at your wrapper script and set optional timeout/retry policies. | Minimal configuration; no code changes required in the AI component. |
| 5️⃣ | **Integrate into CI/Dev workflow**: add the launch command to your local development scripts or to a Homebrew formula for team members. | Makes the tool part of the standard developer environment. |
| 6️⃣ | **Monitor & log**: enable the built‑in logging (--log-file) and add a health‑check endpoint if you need visibility in production. | Helps catch edge cases where the lock is not released. |
| 7️⃣ | **Gradual rollout**: start with internal prototypes, then move to a dedicated Mac mini or on‑prem server for longer‑running jobs. | Reduces risk while you verify stability and maintenance overhead. |

**Production Readiness**

• **Maturity**: Medium. The project is actively maintained (last update 2026‑06‑27) but integration signals are sparse, and documentation is limited to basic usage.
• **Suitability**: Ideal for prototypes, internal tooling, or low‑scale batch jobs where a Mac is already part of the infrastructure. Not recommended as a core component of high‑availability services without additional safeguards (e.g., watchdog scripts, redundancy).
• **Dependencies & Maintenance**: Relies on standard macOS power‑management APIs; no heavy external libraries. Still, you should audit the code for edge‑case handling (e.g., sudden power loss, OS upgrades) and pin a specific release tag.
• **Risk Mitigation**:

1. Verify the repository’s issue tracker for open bugs related to sleep/wake cycles.
2. Add automated tests that simulate agent start/stop to ensure the lid‑lock behaves correctly after OS updates.
3. Consider wrapping the tool in a container‑like sandbox (e.g., brew install + launchd service) to isolate failures.

In summary, *Adrafinil* offers a lightweight, OS‑level solution to keep a closed‑lid Mac awake only while AI agents run, accelerating prototype development. With a modest amount of due‑diligence—license check, functional testing, and monitoring—it can be safely adopted for internal workflows, though it should be treated as a prototype‑grade component rather than a production‑critical service.

Use in a workflow

Show HN: Adrafinil – keep a lid-closed Mac awake only while agents work helps add AI capability without starting from a blank model stack.

• prototype AI features
• build RAG or agent workflows
• evaluate model tooling

Limitations and review notes

Medium: useful for prototypes or internal workflows, with dependency and maintenance checks before production.