接入你的智能体
從發布者安裝。不是 PlugMCP 端點。
遠端 URL 還是本機指令? 兩種安裝方式的差異
Persistent geometric memory for AI agents.
Engram is a local, hardware-native memory substrate that gives AI agents coherent, long-term memory with structure-preserving compression, synthetic calculus over both words and numbers, and true continuity across cold shutdowns.
Share on X / GitHub: docs/images/engram-share-x.png (1280×720). For the repo social preview card: GitHub → Settings → General → Social preview → upload that image.
Unlike vector databases or simple logs, Engram uses fixed-size holographic blocks, VSA operations, sheaf gluing, and categorical reasoning to maintain meaning and relationships even after heavy compression and long-running sessions.
It is designed as a drop-in backend for any LLM (Grok, Claude, Llama, etc.) via the Model Context Protocol (MCP) and is fully open for anyone to build on.
mcp_engram_* tools. Default load set (only two docs): docs/AGENT_MEMORY_CONTRACT.md + docs/skills/engram-wake-up.md. First call every session: mcp_engram_session_start(intent="…"). Do not pre-read five other guides. |Rituals = documented MCP habits (wake → trace decisions → handoff) so memory compounds across sessions — not mysticism, just the discipline that beats flat RAG.
Engram is particularly well-suited for:
Human review (LEG Browser beta): ./scripts/leg (static) or ./scripts/leg --live — see docs/LEG_BROWSER.md.
Full comparison vs mem0/Letta/chroma: see docs/GROK_BUILD_MEMORY.md.
git clone https://github.com/staticroostermedia-arch/engram.git
cd engram
cargo build -p engram-server
target/debug/engram --version # 0.7.0-beta.12
MCP config (Grok Build / Cursor — use scripts/engram-grok):
{
"mcpServers": {
"engram": {
"command": "/path/to/engram/scripts/engram-grok",
"args": ["mcp"],
"env": {
"ENGRAM_STORE": "~/.engram/stalks/",
"ENGRAM_PROFILE": "agent"
}
}
}
}
Restart your IDE, then:
mcp_engram_session_start(intent="your goal")
Lean loop: session_start → context_for_edit(path) → recall(scope=anchors) → quick_trace / remember → session_end(summary).
All ecosystems: integrations/README.md. Cursor ambient wake: ./scripts/cursor-engram-preflight.sh.
Local, read-only mirror of agent memory — no cloud, no npm, no account. Your manifold stays in ~/.engram/; the repo ships tools and the viewer.
./scripts/leg # static — instant curated demo, no backend
./scripts/leg --live # live — engram serve :3456 + viewer :8765
What you get (beta):
session_start)__arc segments, trace chain)Beta caveats: single-file SPA; galaxy view may be slow on 100k+ stores; agent MCP paths stay bounded. Hard-refresh after index.html updates. Static mode is a demo snapshot — --live shows real MCP work.
Full guide: docs/LEG_BROWSER.md. Safe serve restart (does not kill MCP): ./scripts/restart-leg-serve.sh.

Fixed 256KB HolographicBlocks (.leg3): 8192D phase (q), momentum (p), CRS lawfulness, BLAKE3 Merkle, spatial AABB. VSA calculus + sheaf gluing via processes/*.toml (rituals, harness, monitor). NREM / ego.leg3 for long-horizon continuity. Details: docs/GEOMETRIC_MEMORY.md, docs/RITUALS.md, docs/HARNESS_INJECTION.md.
On fast NVMe (e.g. Samsung T700) with cuFile/GPUDirect and full_bvh_gpu recall, the cold manifold on disk acts as a persistent, integrity-checked extension of the agent context window (block footer + verify tools; see CLAIMS_LEDGER.md) — not a separate vector DB you query occasionally.
Each concept is a persistent tensor entry: unit-hypersphere q, momentum p, dynamic bonds (relation blocks + Merkle lineage), CRS ≥ 0.74. The lean path surfaces only the relevant subgraph:
tensor_upsert → relate/bonds → promote_hot → tensor_recall (q preview + edges + lineage)
MCP tools: mcp_engram_tensor_upsert, mcp_engram_tensor_recall (plus lean default recall, query_with_momentum when trending matters). Poll mcp_engram_get_backend_readiness until nvme_recall_ready: true on large stores. Ritual: processes/ritual/solid-tensor-consolidation.toml (p-drift OP_ADD at session_end).
Thought tiles ↔ tensor (unified): mcp_engram_thought_tile_create and write_result dual-write a first-class tensor:tile__{stem} mirror with bonds to goal/trace/spatial concepts. Plain mcp_engram_update on tile:* syncs the mirror; mcp_engram_update_with_tensor_bond is the verified composite. tile_type: propose_improvement routes verified update on target_concept. Wake surfaces rituals via agent_discipline.tensor_unification_rituals. Harness: --suite tensor-thought-unification. Rituals: processes/ritual/thought_tile_to_tensor.toml, verified-update-with-consolidation.toml. See docs/skills/engram-thought-tiles.md and docs/HARNESS_INJECTION.md.
Demo: cargo test -p engram-server solid_state_tensor_verification_harness or examples/tensor_demo.py. Full tile→tensor cycle: STABLE_BIN=target/debug/engram tools/test-harness/bin/engram-harness.sh --suite tensor-thought-unification.
Linguistic calculus (words + numbers in the same sheaf): docs/CATEGORICAL_LINGUISTIC_CALCULUS.md.
flowchart LR
W[session_start<br/>harness injection] --> E[edit + trace]
E --> H[session_end handoff]
H --> W
sig_5 seal + seal-aware lawfulness; wake_digest_v1 + intent-shaped queue; trust residual at wake; Autophagy product GC removed (forget_old = explicit only). See CLAIMS_LEDGER.md.primary_goal rebind on intent mismatch (ENGRAM_PRIMARY_GOAL_REBIND=auto under agent profile); optional ENGRAM_WAKE_DIGEST_ONLY=1 minimal wake packet.ENGRAM_PRAXIS_CONTRACT=hard) when profile is agent (override with soft).ENGRAM_QUALITY_MODE=1 forces eager BVH + readiness flag (CPU stays deferred by default).scope=all dogfood fixes (#209–#213).tool_list() — 101 mcp_engram_* + 4 linguistic); lean default remains 8 essential.to_hybrid_wire / from_hybrid_wire restore q/p/CRS/footer; O_DIRECT .leg3 remains primary on-disk). Continuity spikes / tensor–tile / LEG Browser remain available.Full history: CHANGELOG.md.
Engram supports native synthetic calculus over linguistic structures — including mixed number + word operations — all inside the geometric memory manifold.
Key capabilities:
// Build a linguistic bundle + mixed expression
let bundle = LinguisticDiscourseBundle { ... };
let mixed = op_mixed_linguistic_number_scale(&num_phase, &word);
// Run calculus and store result
let delta = op_linguistic_differentiate(&bundle);
let result = op_linguistic_integrate(&[bundle, delta]);
// Store with full continuity
let _ = Leg3Pointer::mint_linguistic(&result, true); // promotes toward ego.leg3
All operations return CRS (Coherence-Reliability Score) and can be verified with mcp_engram_verify_manifold_integrity.
Build against target/debug/engram during development.
8 essential for daily work — 105 registered (101 mcp_engram_* + 4 linguistic; source: tool_list() in mcp.rs); full map: docs/TOOL_DECISION_MAP.md. Categorized reference: docs/MCP_TOOLS_REFERENCE.md. Cognitive OS: docs/COGNITIVE_OS_EXTENSIONS.md. Harness matrix: tools/test-harness/python/mcp_tool_matrix.py.
Grok plugin slash commands: grok-plugin-engram/commands/.
CLI: engram remember|recall|forget|list|ingest|trace|distill|build-index
Namespaces: mcp_engram_set_namespace("project") or ~/.engram/sheaf.toml
CONTRIBUTING.md · AGENTS.md · PR checklist in .github/PULL_REQUEST_TEMPLATE.md
Dev build: cargo build -p engram-server && target/debug/engram --version
AGPL-3.0-only. .leg3 format: U.S. Patent Application No. 19/372,256 (pending). Commercial licenses: [email protected] — PATENT-NOTICE.md.
Geometric agent memory for Agents: 8-tool lean contract, harness injection at wake, session handoff. 79 MCP tools. Not a vector DB. Rust + MCP.
工具
尚無工具清單。遠端伺服器會在首次健康檢查後取得工具清單。
常見問題
在本頁複製適用於 Cursor、Claude Desktop、Claude Code 或 VS Code 的 stdio 設定。它使用發布者的安裝方式,而不是 PlugMCP 代理伺服器。
這筆記錄未列出必要的環境變數。
Engram 透過 stdio 在本機執行。
上次檢查時間:2026-10-28。 Geometric agent memory for Agents: 8-tool lean contract, harness injection at wake, session handoff. 79 MCP tools. Not a vector DB. Rust + MCP.