DAG-chat – DAG-based AI chat app with branch and merge
Branch and merge conversation paths instead of being stuck in a single linear thread.
1000-line C program where an LLM can edit the DAG mid-run, but only via a four-verb grammar, and every edit is an auditable event
LLM mutates the workflow DAG mid-run via a constrained four-verb grammar.
AI infrastructure engineers, backend developers
LangGraph · Prefect · Airflow
Branch and merge conversation paths instead of being stuck in a single linear thread.
Kubernetes for AI agents with YAML manifests and GitOps workflows.
YAML-to-DAG compilation beats prompt-chaining for reproducible agent workflows.
Metro map metaphor for DAGs is genuinely clever visualization nobody else does.
Multi-session agents with inter-session context passing, but huge feature list hides unclear priorities.
Drag-to-branch is the clear selling point: you can fork any message and pursue alternate prompts while that branch only inherits direct ancestors. The BYOK approach (API key stored in your browser + local Express proxy) keeps it lightweight and safe for a demo, but it's clearly a concept build tied to Claude rather than a multi-model, production-ready tool.