Addresses #875. The previous README was 755 lines mixing six purposes (scam alert, hero, two mea-culpa notes, install guide, architecture explainer, API reference, file map). Rework it as a pure entry point: what MemPalace is, how to install, honest benchmark numbers, links to the website for concept/architecture documentation. Key content changes: - Drop the "highest-scoring AI memory system ever benchmarked" framing. - New tagline: "Local-first AI memory. Verbatim storage, pluggable backend, 96.6% R@5 raw on LongMemEval — zero API calls." Avoids naming a specific vector-store implementation since the backend is pluggable (see mempalace/backends/base.py). - Remove the cross-system comparison table. Retrieval recall (R@5) and end-to-end QA accuracy are different metrics and are not comparable; placing MemPalace's R@5 next to competitor QA accuracy under a single column header was a category error. - The "100%" LongMemEval headline is no longer the lead. The honest held-out figure is 98.4% R@5 on 450 unseen questions. The rerank pipeline reaches >=99% with any capable LLM (reproduced with Claude Haiku, Sonnet, and minimax-m2.7 via Ollama) — pipeline-level, not model-specific. - Benchmark reproduction commands now reference the correct repo (MemPalace/mempalace, not the defunct aya-thekeeper/mempal branch). New file: docs/HISTORY.md as the canonical home for post-launch corrections, public notices, and retractions. Contains verbatim: - 2026-04-14 note on this rewrite (links to #875) - 2026-04-11 impostor-domain notice (moved from README header) - 2026-04-07 "A Note from Milla & Ben" (moved from README body) README keeps a one-line scam-alert callout that links to docs/HISTORY.md for the full timeline.
7.1 KiB
MemPalace — History, Corrections, and Public Notices
This file is the canonical record of post-launch corrections, public notices, and retractions that affect MemPalace's public claims. Newest first.
2026-04-14 — Benchmark table rewrite (issue #875)
A community audit identified a category error in the public benchmark tables
on README.md and mempalaceofficial.com: MemPalace's retrieval recall
numbers (R@5, R@10) were listed in the same columns as competitors'
end-to-end QA accuracy numbers. They are different metrics and are not
comparable — a system can have 100% retrieval recall and 40% QA accuracy.
The audit also found that the retracted "+34% palace boost" claim (see the
April 7 note below) was still present in multiple surfaces despite that
retraction, and that two competitor numbers (Mem0 ~85%, Zep ~85%) had no
published source and did not match the metrics those projects actually
publish.
What changed in this PR:
- The headline number on all surfaces is now 96.6% R@5 on LongMemEval in
raw mode, independently reproduced on Linux x86_64 against the tagged
v3.3.0 release on 2026-04-14. Result JSONLs are committed under
benchmarks/results_*.jsonl(see PR description for the scorecard). - The "100% with Haiku rerank" claim has been removed from all public
comparison tables. It reproduces on our machines and with a different LLM
family (minimax-m2.7 via Ollama Cloud: 99.2% R@5 / 100.0% R@10 on the full
500-question LongMemEval set) — but the 99.4% → 100% step was developed
by inspecting three specific wrong answers (
benchmarks/BENCHMARKS.mdhas called this "teaching to the test" since February). It belongs in the methodology document, not in a headline. - The honest held-out number for the hybrid pipeline — 98.4% R@5 on 450
questions that
hybrid_v4was never tuned on, deterministic seed — is now the comparable figure when an LLM rerank is involved. - The retracted "+34% palace boost" has been removed from
README.md,website/concepts/the-palace.md,website/guide/searching.md, andwebsite/reference/contributing.md. Wing and room filters remain useful — they're standard metadata filters — but they are not presented as a novel retrieval improvement. - Competitor comparison tables mixing retrieval recall with QA accuracy
have been removed from
README.mdandwebsite/reference/benchmarks.md. Where MemPalace can be fairly compared on the same metric, we link to the cited source. Otherwise we report our own numbers and let readers draw their own conclusions. - Reproduction instructions in
benchmarks/BENCHMARKS.mdandbenchmarks/README.mdwere pointing at a defunct branch (aya-thekeeper/mempal); they now point atMemPalace/mempalace. - The LoCoMo 100% R@10 with top-50 rerank row has been removed from
public comparison surfaces. With per-conversation session counts of 19–32
and
top_k=50, the retrieval stage returns every session in the conversation by construction, so the number measures an LLM's reading comprehension over the whole conversation, not retrieval.
Thanks to @dial481 for the detailed audit and to @rohitg00 for the parallel write-up in Discussion #747.
2026-04-11 — Impostor domains and malware
Several community members (issues #267, #326, #506) reported fake MemPalace websites distributing malware. The only official surfaces for this project are:
- This GitHub repository: github.com/MemPalace/mempalace
- The PyPI package: pypi.org/project/mempalace
- The docs site: mempalaceofficial.com
Any other domain — mempalace.tech being the one most commonly reported —
is not ours. Never run install scripts from unofficial sites.
Thanks to our community members for flagging the problem.
2026-04-07 — A Note from Milla & Ben
The community caught real problems in this README within hours of launch and we want to address them directly.
What we got wrong:
The AAAK token example was incorrect. We used a rough heuristic (
len(text)//3) for token counts instead of an actual tokenizer. Real counts via OpenAI's tokenizer: the English example is 66 tokens, the AAAK example is 73. AAAK does not save tokens at small scales — it's designed for repeated entities at scale, and the README example was a bad demonstration of that. We're rewriting it."30x lossless compression" was overstated. AAAK is a lossy abbreviation system (entity codes, sentence truncation). Independent benchmarks show AAAK mode scores 84.2% R@5 vs raw mode's 96.6% on LongMemEval — a 12.4 point regression. The honest framing is: AAAK is an experimental compression layer that trades fidelity for token density, and the 96.6% headline number is from RAW mode, not AAAK.
"+34% palace boost" was misleading. That number compares unfiltered search to wing+room metadata filtering. Metadata filtering is a standard feature of the underlying vector store, not a novel retrieval mechanism. Real and useful, but not a moat.
"Contradiction detection" exists as a separate utility (
fact_checker.py) but is not currently wired into the knowledge graph operations as the README implied."100% with Haiku rerank" is real (we have the result files) but the rerank pipeline is not in the public benchmark scripts. We're adding it.
What's still true and reproducible:
- 96.6% R@5 on LongMemEval in raw mode, on 500 questions, zero API calls — independently reproduced on M2 Ultra in under 5 minutes by @gizmax.
- Local, free, no subscription, no cloud, no data leaving your machine.
- The architecture (wings, rooms, closets, drawers) is real and useful, even if it's not a magical retrieval boost.
What we're doing:
- Rewriting the AAAK example with real tokenizer counts and a scenario where AAAK actually demonstrates compression
- Adding
mode raw / aaak / roomsclearly to the benchmark documentation so the trade-offs are visible- Wiring
fact_checker.pyinto the KG ops so the contradiction detection claim becomes true- Pinning the vector store dependency to a tested range (issue #100), fixing the shell injection in hooks (#110), and addressing the macOS ARM64 segfault (#74)
Thank you to everyone who poked holes in this. Brutal honest criticism is exactly what makes open source work, and it's what we asked for. Special thanks to @panuhorsmalahti, @lhl, @gizmax, and everyone who filed an issue or a PR in the first 48 hours. We're listening, we're fixing, and we'd rather be right than impressive.
— Milla Jovovich & Ben Sigman