io.github.louis49/melchizedek

Official

by louis49 · TypeScript

Persistent memory for Claude Code — offline, single-file, hybrid search.

Persistent offline memory for Claude with hybrid search capability.

Topics

aibm25claude-codedeveloper-toolsembeddingshybrid-searchllmmcpmemorymodel-context-protocolsqlitetypescript

README.md13 KBView on GitHub

Melchizedek

Persistent memory for Claude Code. Automatically indexes every conversation and provides production-grade hybrid search (BM25 + vectors + reranker) via MCP tools. 100% local, zero config, zero API keys, zero invoice.

Why Melchizedek?

Claude Code forgets everything between sessions - and knows nothing about your other projects. Melchizedek fixes both.

It runs silently in the background - indexing your conversations as you work - then gives Claude the ability to search across your entire history, across all projects: past debugging sessions, architectural decisions, error solutions, code patterns.

No cloud. No API keys. No config. Plug and ask.

How it works

code

~/.claude/projects/**/*.jsonl       (your conversation transcripts - read-only)
        |
        v
  SessionEnd hook                   (auto-triggers after each session)
        |
        v
  +-----------------+
  |  Indexer         |    Parse JSONL -> chunk pairs -> SHA-256 dedup
  |  (better-sqlite3)|    FTS5 tokenize -> vector embed (optional)
  +-----------------+
        |
        v
  ~/.melchizedek/memory.db           (single SQLite file, WAL mode)
        |
        v
  +-----------------+
  |  MCP Server      |    16 search & management tools
  |  (stdio)         |    Hybrid: BM25 + vectors + RRF + reranker
  +-----------------+
        |
        v
  Claude Code                       (searches your history via MCP)

Search pipeline - 4 levels of graceful degradation

Every layer is optional. The plugin works with BM25 alone and gets better as more components are available.

Level	Component	What it adds	Dependency
1	BM25 (FTS5)	Keyword search with stemming	None (always active)
2	Dual vectors (sqlite-vec)	Semantic search - text (MiniLM 384d) + code (Jina 768d)	`@huggingface/transformers` (optional)
3	RRF fusion	Merges BM25 + text vectors + code vectors via Reciprocal Rank Fusion	Vectors enabled
4	Reranker	Cross-encoder re-scoring of top results	Transformers.js or node-llama-cpp (optional)

Performance

Measured with npm run bench - 100 sessions, 1 000 chunks, on a single SQLite file.

Metric	Result	Target
Indexation (100 sessions)	~80 ms	< 10 s
BM25 search (mean)	~0.2 ms	< 50 ms
DB size (100 sessions)	~1.4 MB	< 30 MB
Tokens per search	~125	< 2 000

Quick Start

npm (recommended)

bash

npm install -g melchizedek

Add the MCP server to Claude Code:

bash

claude mcp add --scope user melchizedek -- melchizedek-server

npx (no install)

bash

claude mcp add --scope user melchizedek -- npx melchizedek-server

From source

bash

git clone https://github.com/louis49/melchizedek.git
cd melchizedek && npm install && npm run build
claude --mcp-config .mcp.json

Claude Code plugin marketplace (coming soon)

Plugin review pending. In the meantime, use npm or npx install above.

bash

claude plugin install melchizedek   # not yet available

Setting up hooks (automatic indexing)

The MCP server provides search tools, but hooks trigger automatic indexing. Without hooks, you'd need to manually index sessions.

For marketplace installs, hooks are configured automatically. For npm/npx/source installs, add hooks to ~/.claude/settings.json.

See docs/installation.md for the full JSON configuration, hook reference, and troubleshooting.

After setup, restart Claude Code. Indexing starts automatically.

MCP Tools

Search (start here)

Tool	Description
`m9k_search`	Search indexed conversations. Returns compact snippets. Current project boosted. Supports `since`/`until` date filters and `order` (score, date_asc, date_desc).
`m9k_context`	Get a chunk with surrounding context (adjacent chunks in the same session).
`m9k_full`	Retrieve full content of chunks by IDs.

Progressive retrieval pattern - search returns ~50 tokens/result, context ~200-300, full ~500-1000. Start with m9k_search, drill down only when needed. 4x token savings vs loading everything.

Context-aware ranking - results from your current project (×1.5) and current session (×1.2) are automatically promoted. Cross-project results remain visible.

Specialized search

Tool	Description
`m9k_file_history`	Find past conversations that touched a specific file.
`m9k_errors`	Find past solutions for an error message.
`m9k_similar_work`	Find past approaches to similar tasks. Prioritizes rich metadata.

Memory management

Tool	Description
`m9k_save`	Manually save a memory note for future recall.
`m9k_sessions`	List all indexed sessions, optionally filtered by project.
`m9k_info`	Show memory index info: corpus size, search pipeline, embedding worker, usage metrics.
`m9k_config`	View or update plugin configuration.
`m9k_forget`	Permanently remove a chunk from the index.
`m9k_delete_session`	Delete a session from the index.
`m9k_ignore_project`	Exclude a project from indexing. Future sessions won't be indexed, existing ones optionally purged.
`m9k_unignore_project`	Re-enable indexing for a previously ignored project. Purged data is not restored.
`m9k_restart`	Restart the MCP server to load fresh code after `npm run build`. Supports `force: true` for stuck processes.

Usage guide

Tool	Description
`__USAGE_GUIDE`	Phantom tool. Its description teaches Claude the retrieval pattern and available tools.

Configuration

Zero config by default. Everything is tunable via m9k_config or environment variables.

Setting	Default	Env var
Database path	`~/.melchizedek/memory.db`	`M9K_DB_PATH`
Daemon mode	enabled	`M9K_NO_DAEMON=1` to disable
Log level	`warn`	`M9K_LOG_LEVEL`
Embeddings enabled	`true`	`M9K_EMBEDDINGS=false` to disable
Reranker enabled	`true`	`M9K_RERANKER=false` to disable

See docs/configuration.md for the full settings reference (20+ options, env vars, config file examples).

Enhanced Search

Melchizedek works out of the box with BM25 keyword search. Text embeddings (MiniLM) download automatically on first use for semantic search.

For GPU-accelerated code embeddings (Ollama), cross-encoder reranking (GGUF models), platform-specific setup guides, and the full model reference, see Enhanced Search Setup.

How is this different?

	Melchizedek	claude-historian-mcp	claude-mem	episodic-memory	mcp-memory-service

Philosophy	Search engine - indexes everything, you search	Search engine - scans JSONL on demand	Notebook - AI compresses & saves	Search engine	Notebook - AI decides what to store
Indexes raw conversations	Yes (JSONL transcripts)	Yes (direct JSONL read, no persistent index)	Compressed summaries	Yes (JSONL)	No (manual `store_memory`)
Retroactive on install	Yes (backfills all history)	Yes (reads existing files)	No	Yes	No (empty at start)
Search	BM25 + vectors + RRF + reranker	TF-IDF + fuzzy matching	FTS5 + ChromaDB	Vectors only	BM25 + vectors
Progressive retrieval	3 layers (search/context/full)	No	No	No	No
100% offline	Yes	Yes	No (needs API for compression)	Yes	Yes
Single-file storage	SQLite	None (reads raw JSONL)	SQLite + ChromaDB	SQLite	SQLite-vec
Zero config	Yes	Yes	Yes	Yes	Yes
MCP tools	16	10	4	2	12
License	MIT	MIT	AGPL-3.0	MIT	Apache-2.0
Dual embedding (text + code)	Yes (MiniLM + Jina Code)	No	No	No	No
Configurable models	Yes (Transformers.js or Ollama)	No	No (Chroma internal)	No (hardcoded)	Yes (ONNX, Ollama, OpenAI, Cloudflare)
Reranker	Cross-encoder (ONNX, GGUF, or HTTP)	No	No	No	Quality scorer (not search reranker)
Privacy	All local, `<private>` tag redaction	All local	Sends data to Anthropic API	All local	All local
Multi-instance	Singleton daemon - N Claude windows share 1 process (Unix socket / Windows named pipe, local fallback)	N separate processes	Shared HTTP worker (:37777)	N separate processes	Shared HTTP server

Inspirations

This project stands on the shoulders of others. Key ideas borrowed from:

Project	What we took
CASS	RRF hybrid fusion, SHA-256 dedup, auto-fuzzy fallback
claude-historian-mcp	Specialized MCP tools (file_history, error_solutions)
claude-diary	PreCompact hook (archive before `/compact`)

Known issues

Session boost inactive - Claude Code currently sends an empty session_id in the SessionStart hook stdin payload, preventing the ×1.2 session boost from working. The ×1.5 project boost is unaffected and provides the primary context-aware ranking. Related upstream issues: #13668 (empty transcript_path), #9188 (stale session_id). Melchizedek's session boost code is tested and ready, and will activate automatically when the upstream fix lands.

Privacy

Zero telemetry. No tracking, no analytics, no network calls (except optional lazy model download).
Read-only on transcripts. Never writes to ~/.claude/projects/. All data in ~/.melchizedek/.
<private> tag support. Content between <private>...</private> is replaced with [REDACTED] before indexing.
Local-only. Your conversations never leave your machine.

Requirements

Node.js >= 20
Claude Code >= 2.0
macOS, Linux, or Windows

License

MIT

"Without father, without mother, without genealogy, having neither beginning of days nor end of life."

Hebrews 7:3

Built by @louis49

Install

claude_desktop_config.json

{
  "mcpServers": {
    "melchizedek": {
      "command": "npx",
      "args": [
        "-y",
        "melchizedek@1.0.2"
      ]
    }
  }
}