ActiveGraph: An Event-Sourced Runtime for Auditable, Forkable Agentic Systems
By
[Submitted on 21 May 2026]
Summary
ActiveGraph is a runtime architecture that inverts the typical agent framework design by making the append-only event log the source of truth, rather than bolting logging onto a language-model-first loop. The working graph is a deterministic projection of the event log, and behaviors (functions, classes, LLM routines) react to graph changes and emit new events — with no component directly instructing another. This yields three key properties: deterministic replay from the log, cheap forking of runs without re-executing shared prefixes, and end-to-end lineage from high-level goals to individual model calls. The paper presents the architecture, a determinism contract for sound replay, and a worked example whose causal structure is reconstructable from the log alone.
Source
Hacker NewsActiveGraph: An Event-Sourced Runtime for Auditable, Forkable Agentic Systemsarxiv.orgKey quotes
· 4 pulledThe append-only event log is the source of truth; the working graph is a deterministic projection of that log.
No component instructs another; coordination happens entirely through the shared graph.
This single design decision yields three properties that retrieval-and-summarization memory systems do not provide: deterministic replay of any run from its log, cheap forking that branches a run at any event without re-executing the shared prefix, and end-to-end lineage from a high-level goal down to the individual model call that produced each artifact.
Most agent frameworks are built around the language model: a conversation loop comes first, then tools, then rules, and finally a logging layer bolted on for observability, with state persisted as retrievable 'memory.'
You might also wanna read
Autonomous Agentic Systems: A Practical Guide to "Always-On" Agents
Agora: Toward Autonomous Bug Detection in Production-Level Consensus Protocols with LLM Agents
Arbor: A Multi-Agent Framework Using Tree Search for Autonomous LLM Inference Optimization
Arbor is a multi-agent framework that uses structured tree search as a cognition layer for autonomous agents operating in large, stateful ac
arxiv.org·20d ago
AWS Builder Center: Implementing an Agentic Planner-Generator-Evaluator Pattern with Memory Using LangGraph
This article, published at AWS Builder Center, describes an advanced implementation of an Agentic Planner-Generator-Evaluator pattern using
dev.to·20d ago
Meta Tech Lead Nishant Gupta Makes Case for Deterministic Infrastructure to Support Non-Deterministic AI Agents
Nishant Gupta, a Tech Lead at Meta, presents on the critical need for deterministic infrastructure to support non-deterministic AI agents in
startuphub.ai·6d ago
Memori launches agent-native persistent memory infrastructure using structured knowledge graphs from agent trace data
Memori is a new agent-native memory infrastructure that enables AI agents to create structured, long-term persistent memory directly from ag
Product Hunt·1mo ago
Comments
Sign in to join the conversation.
No comments yet. Be the first.