Back to Tool
E

engrams

Tool Oct 19, 2026

A self-hosted orchestrator that runs Claude Code and Codex sessions in Firecracker microVMs, with snapshots, automations and a dashboard.

Visit primary link 3 sources attached
engrams dashboard shows an agent conversation, a shared generated picture and a shell pane with the program used to produce it.
Maintainer-published session screenshot from the versioned engrams README. The example output and tool activity illustrate the existing dashboard; they are not an AITrending test or evidence of a successful cross-host migration. Source: Cortex Applications / engrams contributors
agent-ecosystemcoding-agentsself-hosted

What happened lately

engrams timeline

engrams is infrastructure for running coding agents in an operator’s own cloud. It provides a dashboard, CLI, Slack integration and trigger-based automations around separate agent sessions. Claude Code and Codex are bundled harnesses; engrams itself does not supply a model. Users configure provider credentials and pay the relevant model and infrastructure costs.

Production uses Linux/KVM Firecracker microVMs on a dedicated Kubernetes host fleet, with Terraform and Helm deployment paths for GKE and EKS. A filtering proxy limits session network destinations. The local development stack also supports Apple Silicon virtualization and an unisolated subprocess fallback; those modes are not interchangeable with production isolation.

The main project uses AGPL v3, while specified harness SDK crates use Apache 2.0. Version 0.11.0 adds durable cross-host session moves and harness event continuity, according to its maintainers. It remains a 0.x project with breaking upgrade requirements, rather than a hosted sandbox service or a model provider.