The attack is a sleeping sentence your logs cannot see

Vestige, a local-first memory layer for AI agents, launches as a 25MB Rust binary over MCP that uses causal and temporal links to trace failures back to their root causes, citing research from Nature 2024 and DeepMind's arXiv paper. The tool, which requires Node.js and works with Claude Code, Codex, Cursor, and others, claims to detect contradictions, merge redundant memories, and fade unused ones, with all data staying on the user's machine.

Local-first memory for AI agents that finds the cause, not just the match. Vestige remembers your decisions, catches contradictions before they cost you, and traces a failure back to the older memory that actually caused it. One 25MB Rust binary over MCP. No cloud, no API keys, no telemetry. Your data never leaves your machine. Install install · Why not RAG why-not-just-rag · Benchmark the-receipts-silent-rotation · Science the-science · Tools the-14-tools · Dashboard the-dashboard · Pro vestige-pro · Docs go-deeper Agents re-learn the same lessons: they recommend a change you already tested and rejected, re-derive a fix that was already written down, and treat every session as if the last one never happened. Vestige is the memory layer that ends that. Any MCP-capable agent Claude Code, Claude Desktop, Codex, Cursor, and others writes memories as you work and retrieves them later, modeled on real cognitive science: redundant memories merge, contradicted ones are flagged, unused ones fade, and when a failure hits, Vestige reaches backward to the decision that set it up. The cause never looks like the bug. That is the whole product. You need Node.js. No Docker, no signup, no compile step prebuilt for macOS ARM + Intel, Linux x86 64, Windows x86 64 . npm install -g vestige-mcp-server@latest Connect it to your agent. Every MCP client understands this config: { "mcpServers": { "vestige": { "command": "vestige-mcp" } } } | Client | Setup | |---|---| | Claude Code | claude mcp add vestige vestige-mcp -s user | | Codex | codex mcp add vestige -- vestige-mcp | | Cursor / VS Code / Windsurf | | docs/CONFIGURATION.md /samvallad33/vestige/blob/main/docs/CONFIGURATION.md claude-desktop-macos Verify: vestige dashboard , then open http://localhost:3927/dashboard . First run downloads a 130MB embedding model once; after that Vestige is fully offline, forever. Full walkthrough: docs/GETTING-STARTED.md /samvallad33/vestige/blob/main/docs/GETTING-STARTED.md . RAG retrieves text that resembles the query. That is the right tool when the answer looks like the question, and the wrong tool when the cause of a problem looks nothing like the symptom: a config choice from three weeks ago, a library pin, an assumption nobody flagged as risky. | Vector search | Vestige | | |---|---|---| | Retrieval basis | Similarity to the query | Causal + temporal links, plus similarity | | Root cause of a failure | Cannot; the cause does not resemble the bug | vestige backfill --contrast reaches backward to it | | Contradictions | Both stored, both returned | Detected and flagged claim contradicts memory | | Redundant writes | Accumulate | Merged on write prediction-error gating | | Unused memories | Persist at full weight | Fade FSRS-6 spaced repetition | | Your data | Usually a cloud service | Never leaves your machine | The backward reach implements Retroactive Salience Backfill Zaki, Cai et al., Nature 2024, 637:145-155, DOI 10.1038/s41586-024-08168-4 https://doi.org/10.1038/s41586-024-08168-4 : when a memory turns out to matter, the salience of the earlier memories that led to it is raised, so the causal chain becomes retrievable even though the surface text never matched. Every backfill result ships with a receipt naming the exact evidence path; Vestige reports receipt-backed candidate causes, never an unverifiable verdict. And the limitation on the left column is not marketing: DeepMind proved single-vector retrieval mathematically incapable of certain relevance patterns arXiv:2508.21038 https://arxiv.org/abs/2508.21038 , ICLR 2026 . The claim is testable, and the test ships with all 246 agent transcripts it produced. Three coding agents fix one failing e2e test; the fix needs the currently live signing key id, randomized per trial from a 50-key keyring, present in no file the agents can read. It exists only in the memory layer. The dangerous outcome is converging on a planted decoy: tests pass, the merge is clean, production breaks. | Arm 6 models, 25 trials | Converged correct | Converged wrong | Split | |---|---|---|---| | No memory | 0/25 | 21/25 | 4/25 | | Dense cosine RAG | 4/23 | 12/23 | 7/23 | | Vestige | 20/23 | 0/23 | 3/23 | On the verbatim queries the agents typed, the causal memory ranks 7th of 8 under both dense cosine and BM25 while the decoy ranks 1st. Reproduce the central measurement in two seconds, stdlib only: git clone -b benchmark/silent-rotation --depth 1 https://github.com/samvallad33/vestige.git cd vestige/benchmarks/silent-rotation python3 tests/bm25 baseline.py results/runA-trial-1/corpus-export.json --no-dense The caveats are published alongside the results, including the trials a plain cosine baseline ties and the trial Vestige loses. Every mechanism is a cited result, implemented in Rust, running locally. Full write-up: docs/SCIENCE.md /samvallad33/vestige/blob/main/docs/SCIENCE.md . | Mechanism | What it does | Source | |---|---|---| | Prediction-Error Gating | Stores only the novel; merges redundant, flags contradictory | Hippocampal novelty gating | | FSRS-6 spaced repetition | Used memories persist, unused ones fade | Modern spaced-repetition research | | Retroactive Salience Backfill | Reaches backward to a failure's root-cause memory | Zaki, Cai et al. 2024, Nature | | Synaptic Tagging | Marks memories for later consolidation | Frey & Morris 1997 | | Spreading Activation | One retrieval activates related memories through the graph | Collins & Loftus 1975 | | Dual-Strength | Storage strength vs retrieval strength, tracked separately | Bjork & Bjork 1992 | | Memory Dreaming | Sleep-like replay and synthesis | Sleep consolidation research | | Active Forgetting | Reversible top-down suppression, cascading to neighbors | Anderson 2025, Davis 2020 | Your agent calls these; you rarely do. | Tool | Purpose | |---|---| recall | Retrieve memories relevant to the current context | smart ingest | Store a fact, gated for novelty and contradiction | backfill | Reach backward from a failure to its candidate cause | receipt | Inspect retrieval receipts and evidence replay | memory · graph · intention maintain · dedup · suppress memory status · codebase · source sync · session start Project scoping, hygiene workflows, and making memory a standing habit for your agent: docs/MEMORY HYGIENE.md /samvallad33/vestige/blob/main/docs/MEMORY HYGIENE.md · docs/AGENT-MEMORY-PROTOCOL.md /samvallad33/vestige/blob/main/docs/AGENT-MEMORY-PROTOCOL.md · docs/CLAUDE-SETUP.md /samvallad33/vestige/blob/main/docs/CLAUDE-SETUP.md . vestige dashboard A living WebGPU observatory of your memory at http://localhost:3927/dashboard : memories appear, link, strengthen, and fade in real time, 1000+ nodes at 60fps. It renders a deterministic 12-second loop of your store's life that you can export as an mp4 with one click, and mints a brain print , a signature seeded from your store's shape. Share artifacts are structure-only by design: your brain, never your memories. Everything above is free forever and never metered. Pro $19/month is managed, end-to-end encrypted continuity: your memory graph and accountability history receipts, traces, memory PRs following you across machines. XChaCha20-Poly1305 applied on your device, Argon2id over a passphrase only you know, ciphertext-only server. Zero-knowledge is the design: lose the passphrase and the data is unrecoverable, by anyone. Checkout opens shortly; watch Releases https://github.com/samvallad33/vestige/releases for the announcement. | Engine | Rust 2024, ~96k lines, single 25MB binary, 1,961 tests, clippy clean at -D warnings | | Retrieval | Nomic Embed v1.5 Matryoshka 768d→256d + USearch HNSW + SQLite FTS5, optional Qwen3 reranker | | Storage | SQLite, optional SQLCipher encryption | Getting Started /samvallad33/vestige/blob/main/docs/GETTING-STARTED.md · FAQ /samvallad33/vestige/blob/main/docs/FAQ.md · The Science /samvallad33/vestige/blob/main/docs/SCIENCE.md · Configuration /samvallad33/vestige/blob/main/docs/CONFIGURATION.md · Storage /samvallad33/vestige/blob/main/docs/STORAGE.md · Silent Rotation https://github.com/samvallad33/vestige/tree/benchmark/silent-rotation/benchmarks/silent-rotation · Changelog /samvallad33/vestige/blob/main/CHANGELOG.md If Vestige saves you from one repeated mistake, that is the whole point: never solve the same problem twice. If it earns a place in your setup, a star https://github.com/samvallad33/vestige genuinely helps.