Python API

Auto-generated reference for the mempalace Python package. Source of truth lives in the docstrings under mempalace/ — edit there, not here. Regenerate with scripts/render-api-docs.py.

For task-oriented overviews of the main interfaces (search, memory stack, knowledge graph, palace graph, AAAK dialect, configuration), see Python API Overview.

Modules

Top-level modules

• mempalace.auto_query — Auto-query integration for MemPalace.
• mempalace.auto_wake — Wake-on-demand for a sleeping palace host.
• mempalace.backends — Storage backend implementations for MemPalace (RFC 001).
• mempalace.backfill_age — Backfill the AGE graph from an existing drawer table.
• mempalace.backups — Writing and pruning palace backups.
• mempalace.calibration — Calibrated confidence for search results.
• mempalace.cli — MemPalace — Give your AI a memory. No API key required.
• mempalace.closet_llm — closet_llm.py — Generate closets via a user-configured LLM for richer indexing.
• mempalace.collision_scan — Pre-mining defense against drawer_id collisions.
• mempalace.config — MemPalace configuration system.
• mempalace.convo_miner — convo_miner.py — Mine conversations into the palace.
• mempalace.convo_scanner — convo_scanner.py — Parse Claude Code conversation directories into ProjectInfo.
• mempalace.corpus_origin — corpus_origin.py — Detect whether a corpus is an AI-dialogue record and,
• mempalace.cross_encoder_rerank — Optional cross-encoder reranking for the retrieval path.
• mempalace.daemon — Long-lived local daemon for queued MemPalace writes.
• mempalace.date_window — Shared date-window parsing for read-path filters.
• mempalace.dedup — dedup.py — Detect and remove near-duplicate drawers
• mempalace.dialect — AAAK Dialect -- Structured Symbolic Summary Format
• mempalace.diary_ingest — diary_ingest.py — Ingest daily summary files into the palace.
• mempalace.dynamics — dynamics.py — Living-connection math for halls + tunnels.
• mempalace.embedding — Embedding function factory with hardware acceleration.
• mempalace.encoding_repair — Safely repair high-confidence UTF-8 mojibake in MemPalace drawers.
• mempalace.entities — No-LLM structural entity extraction for the associative graph.
• mempalace.entity_detector — entity_detector.py — Auto-detect people and projects from file content.
• mempalace.entity_registry — entity_registry.py — Persistent personal entity registry for MemPalace.
• mempalace.exporter — exporter.py — Export the palace as a browsable folder of markdown files.
• mempalace.fact_checker — fact_checker.py — Verify text against known facts in the palace.
• mempalace.format_miner — format_miner.py — proposed for mempalace 3.3.6.
• mempalace.general_extractor — general_extractor.py — Extract 5 types of memories from text.
• mempalace.hallways — Hallways — within-wing entity-to-entity connectors.
• mempalace.hlc — hlc.py — Hybrid Logical Clock for RFC 004 op ordering
• mempalace.hook_shell — Compatibility helpers for legacy shell hooks.
• mempalace.hooks_cli — Hook logic for MemPalace — Python implementation of session-start, stop, session-end, and precompact hooks.
• mempalace.ids — Centralized drawer/triple ID construction with collision-safe delimiter.
• mempalace.instructions_cli — Instruction text output for MemPalace CLI commands.
• mempalace.kg_canonical_vocab — Closed-vocabulary predicate mapping spike (issue #72).
• mempalace.kg_canonical_writepass — Guarded post-extraction canonical-mapping write pass (issue #72, approach a).
• mempalace.kg_llm_extractor — LLM-based triple extractor for the async KG worker.
• mempalace.kg_predicate_norm — Predicate normalization for the AGE knowledge graph (issue #50).
• mempalace.kg_triple_worker — Async worker that drains mempalace_kg_extraction_queue.
• mempalace.kg_writethrough — KG write-through hooks for PostgresCollection drawer writes.
• mempalace.knowledge_graph — knowledge_graph.py — Temporal Entity-Relationship Graph for MemPalace
• mempalace.knowledge_graph_age — AGE-backed implementation of KnowledgeGraph (Apache AGE on Postgres).
• mempalace.layers — layers.py — 4-Layer Memory Stack for mempalace
• mempalace.llm_client — llm_client.py — Minimal provider abstraction for LLM-assisted entity refinement.
• mempalace.llm_refine — llm_refine.py — Optional LLM refinement of regex-detected entities.
• mempalace.logstream — logstream.py — Agent coordination event log for MemPalace (RFC 003)
• mempalace.logsync — logsync.py — Anti-entropy sync engine for the logstream (RFC 004 step 0)
• mempalace.mcp_proxy — Thin stdio front end for the MemPalace MCP server.
• mempalace.mcp_server — MemPalace MCP Server — read/write palace access for Claude Code
• mempalace.migrate — mempalace migrate — Recover a palace created with a different ChromaDB version.
• mempalace.migrate_to_postgres — ChromaDB → Postgres (pgvector + AGE) migration tool.
• mempalace.miner — miner.py — Files everything into the palace.
• mempalace.multi_encoder — Multi-encoder retrieval — query N encoder-bound palaces, RRF-fuse.
• mempalace.normalize — normalize.py — Convert any chat export format to MemPalace transcript format.
• mempalace.novelty — novelty.py — Gzip-based novelty scoring for drawers
• mempalace.novelty_wiring — novelty_wiring.py — write-time novelty tagging
• mempalace.onboarding — onboarding.py — MemPalace first-run setup.
• mempalace.palace — palace.py — Shared palace operations.
• mempalace.palace_graph — palace_graph.py — Graph traversal layer for MemPalace
• mempalace.palace_graph_age — Palace structure (Wing → Room → Drawer) as native AGE graph nodes.
• mempalace.pending_queue — Append-only journal for mine requests that couldn't reach the daemon.
• mempalace.project_scanner — project_scanner.py — Detect projects and people from real signal.
• mempalace.query_sanitizer — query_sanitizer.py — Mitigate system prompt contamination in search queries.
• mempalace.ratings — Feedback ratings for search results (#159, Tier 1).
• mempalace.recency — Recency weighting for search results (#158).
• mempalace.repair — repair.py — Scan, prune corrupt entries, and rebuild HNSW index
• mempalace.replica — replica.py — Per-palace replica identity (RFC 004 transport seam / provenance)
• mempalace.room_detector_local — room_detector_local.py — Local setup, no API required.
• mempalace.room_taxonomy — Canonical room taxonomy — soft-warn validation.
• mempalace.rrf — Reciprocal Rank Fusion — combine ranked lists from N retrievers.
• mempalace.searcher — searcher.py — Find anything. Exact words.
• mempalace.server_registry — Per-palace discovery registry for the MemPalace HTTP MCP hub.
• mempalace.service — Shared service operations used by daemon-backed entry points.
• mempalace.sources — Source adapter subsystem (RFC 002).
• mempalace.spellcheck — spellcheck.py — Spell-correct user messages before palace filing.
• mempalace.split_mega_files — split_mega_files.py — Split concatenated transcript files into per-session files
• mempalace.sweeper — sweeper.py — Message-granular miner that catches what the file-level
• mempalace.sync — sync.py — Gitignore-aware drawer prune (#1252).
• mempalace.tag_extraction — TF-IDF auto-tag extraction for drawer write time (#201).
• mempalace.tags — Multi-label tags for drawers (techempower-org/mempalace#39).
• mempalace.transport — transport.py — The RFC 004 transport seam (Layer 1).
• mempalace.version — Single source of truth for the MemPalace package version.
• mempalace.wal — Side-effect-free write-ahead log for MemPalace write operations.
• mempalace.write_routing — Shared daemon-routing policy for MemPalace write callers.
• mempalace.write_sanitizer — write_sanitizer.py — Observation-grade input hygiene for the write path (#40).

auto_query

• mempalace.auto_query.decisions — Decision logger for the auto-query system.
• mempalace.auto_query.depth_cache — TTL cache for the periodic depth-refresh injection.
• mempalace.auto_query.formatter — Result formatter for the auto-query integration.
• mempalace.auto_query.router — Tool router for auto-query integration.
• mempalace.auto_query.runner — Auto-query runner — chains signal extraction, routing, and formatting.
• mempalace.auto_query.signals — Signal extraction for auto-query context classifier.

backends

• mempalace.backends._sidecar — Shared embedder-identity sidecar (RFC 001).
• mempalace.backends.base — Storage backend contract for MemPalace (RFC 001).
• mempalace.backends.chroma — ChromaDB-backed MemPalace storage backend (RFC 001 reference implementation).
• mempalace.backends.embedding_wrapper — Core-side embedding adapter for explicit-vector backends.
• mempalace.backends.milvus — Milvus backend for MemPalace.
• mempalace.backends.pgvector — Postgres + pgvector backend for MemPalace.
• mempalace.backends.postgres — Optional PostgreSQL-backed MemPalace storage backend.
• mempalace.backends.qdrant — Qdrant REST backend for MemPalace.
• mempalace.backends.registry — Backend registry + entry-point discovery (RFC 001 §3).
• mempalace.backends.sqlite_exact — SQLite exact-vector backend for MemPalace.

integrations

• mempalace.integrations.hermes — MemPalace memory provider for Hermes.

sources

• mempalace.sources.aider — Aider source adapter (RFC 002).
• mempalace.sources.base — Source adapter contract for MemPalace (RFC 002).
• mempalace.sources.codex — Codex CLI source adapter (RFC 002).
• mempalace.sources.context — PalaceContext facade passed to source adapters (RFC 002 §9).
• mempalace.sources.conversations — Conversation source adapter (RFC 002 §9).
• mempalace.sources.filesystem — Filesystem source adapter (RFC 002 §9).
• mempalace.sources.gemini — Gemini CLI source adapter (RFC 002).
• mempalace.sources.opencode — OpenCode source adapter (RFC 002).
• mempalace.sources.registry — Source adapter registry + entry-point discovery (RFC 002 §3).
• mempalace.sources.transforms — Reference implementations of the reserved content transformations (RFC 002 §1.4).
• mempalace.sources.warp — Warp terminal source adapter.