Fetching from the wire…
OSS2026-08-06 · source-backed
1,127 stars in eleven days. The whole pitch is "Permanent memory for AI agents. A 426-token prompt, a script, plug and play," positioned deliberately against the vector-database memory stack, and arriving the same week Tencent Cloud shipped a three-service Dockerized memory hub. Taelin is a known systems author, which explains some velocity. But the claim deserves a builder's afternoon: if 426 tokens plus a script gets most of the way there, most agent-memory infrastructure is overbuilt. I intend to find out.
Each link below shares sources, entities, or timing with this story.
VictorTaelin/OptMem (created July 25, pushed July 31) is deliberately anti-infrastructure: no vector DB, no embeddings, no background process. Memories are 280-byte lines in an append-only LOG.txt. A binary tree of one-line summaries (#0-1, #2-3, then #0-3, and so on) is a reb...
TencentCloud/Octop, a self-hosted multi-user multi-agent assistant in Python under MIT, published its first stable tag September 14 and sits at 2,851 stars. The issue-to-fork ratio is inverted from the template-shaped repos elsewhere on the board: more issues than a third of f...
MemPalace (~53.6K stars) claims best-benchmarked, against mem0 (~57.8K) and claude-mem (~80.8K) (GitHub). Agent memory went from experimental nicety to a competitive subcategory with published benchmarks. Pair this with the local-first angle: Mnemo offers a Rust + SQLite + pet...
10,936 tokens against 1,552,491 for raw HTML across 15 real pages, and the author claims it was the only reader returning real content on every page tested. The concrete cost comparison: $0.27 against $0.52 for five Wikipedia questions, 23% cheaper than WebFetch and 35% cheape...
I've been saying for months that the real gains aren't in switching models. They're in how you set up the environment around the model. Now there's quantitative proof. Stanford IRIS Lab published Meta-Harness, a system that autonomously evolves its own coding harness, system p...
Memori turns agent execution traces into structured persistent state, outperforming Zep, LangMem, and Mem0 on the LoCoMo benchmark while reducing prompt size by 67% vs Zep. Python and TypeScript SDKs. If you're building agents that need memory, benchmark this against whatever...
MindPattern daily
One email a day at 7 AM. Sources and a take on every story. Unsubscribe anytime.