Orchestrating autonomous coding agents

I built Aergia to answer a simple question: what happens if you let coding agents run unattended, around the clock, with real isolation and real memory? Here are the parts that mattered.

One project, one box

Every project gets its own Docker container. Agents are powerful and occasionally wrong; isolation turns "wrong" into "contained." A blast radius of one repository is the difference between a useful tool and a liability.

// each run is sandboxed and disposable
const box = await pool.lease(project.id);
await box.run(agent, { memory: project.memory, gitToken: project.token });

Memory that survives the session

A stateless agent re-learns the codebase every run. Persistent, per-project memory — what was tried, what worked, what's off-limits — is what turns a sequence of one-shot prompts into something that compounds.

Keeping it honest

The failure mode of 24/7 autonomy isn't the dramatic one; it's quiet drift. Strict CI quality gates, a human-readable activity log, and hard stops on anything touching production keep the system trustworthy when nobody's watching.

The result runs as a TypeScript monorepo on self-hosted infrastructure, driven from Telegram and a real-time dashboard. It's less "magic" and more "good operations" — which is exactly why it works.