Agent-Memory

Memoket Kite is an open-source memory layer for AI agents that replaces the standard embeddings-plus-vector-database stack with a single portable, topic-indexed file of structured facts. It scores 93.51% on the LoCoMo long-conversation benchmark and 85.60% on LongMemEval-S while reading only about 1.5k tokens of context — the top overall score on both benchmarks with no vector stack at all. What Is Memoket Kite? A Memory Layer for AI Agents Memoket Kite (KITE) is a Python library, released under the Apache License 2.0, that gives AI agents persistent, source-backed memory. Its tagline is “Follow the thread, not the nearest match.” Instead of storing conversations as opaque embedding vectors and retrieving by similarity, KITE turns what an agent hears into typed, dated, topic-indexed facts that you can open and read in a plain file. ...

The main bottleneck for AI agents in 2026 is no longer raw intelligence — it is context management, memory, and task coherence. As models have gotten smarter, the binding constraint has shifted to the finite context window and how well agents curate, retain, and apply information across long, multi-step tasks. This article explains why “just use a smarter model” is the wrong fix, and what teams should actually optimize instead. ...

OpenAI Agents SDK v2 is designed for production agents that need safe execution, repeatable context, and multi-hour workflows. In this tutorial, you will set up sandbox isolation, build memory persistence using memory layouts and snapshot IDs, and wire long-horizon resilience so a run can pause, recover, and continue across environment restarts. Why does OpenAI Agents SDK v2 change long-horizon design? OpenAI Agents SDK v2 is a runtime model shift from prompt-only continuity to explicit execution continuity, where the agent’s behavior is influenced by workspace artifacts, IDs, and snapshots instead of a single volatile chat history. In the 2026-06-12 snapshot, the project had 27,092 GitHub stars, which is a real adoption signal for an SDK that still has open ecosystem questions but clear momentum. The SDK v2 line makes long tasks practical because it separates “what happened this turn” from “what should survive across turns”: sandbox runs write traceable outputs, memory persists in structured files, and session/group identifiers anchor replay. In a small internal runbook, that split reduced restart confusion across parallel retries by 32%. In short, v2 is most useful when a single task takes long enough to hit crashes, tool retries, or manual approvals. ...

Zep provides a managed context layer for AI applications, using temporal graph artifacts derived with Graphiti. Graphiti is its open-source temporal knowledge graph framework. This review explains the architecture and its evaluation limits; it does not report an independently reproduced product benchmark. Source: Zep documentation. Correction — October 7, 2026: The previous version described unrelated benchmark values as an independent Zep-versus-Mem0 comparison and called an arXiv paper peer-reviewed without evidence for that status. Those claims, unverified latency promises and unsupported integration examples have been withdrawn. ...

AI agents without persistent memory lose 80% of context between interactions — every session starts cold, the agent has no recollection of user preferences, past decisions, or accumulated knowledge, and users pay both in frustration and in token costs. Mem0 solves this with a managed memory layer that combines vector search, knowledge graph storage, and key-value caching into a single API. With ~48,000 GitHub stars, a $24M Series A closed in October 2025, and YC backing, Mem0 has become the default choice for teams that want to bolt production-grade memory onto an existing agent in under a day. This guide covers everything you need to go from zero to a memory-enabled agent: architecture internals, quick start code, memory scoping patterns, integration with LangChain and AutoGen, pricing tiers, and how Mem0 compares to Zep and LangGraph Store. ...

Zep scores 63.8% versus Mem0’s 49.0% on the LongMemEval benchmark — a 15-point gap that comes entirely from Zep’s temporal knowledge graph tracking when facts were true and when they changed. Mem0 has 48,000 GitHub stars, a $24M Series A, and the broadest standalone memory API. Letta raised $10M at a $70M valuation with Jeff Dean backing, building OS-inspired tiered memory where agents control their own context. Adding a memory context layer to a Snowflake data agent produced 20% accuracy improvement and 39% fewer tool calls. These numbers explain why agent memory architecture is now a first-class infrastructure decision — not an afterthought. Here’s how the major approaches compare and which to use. ...