Octobrain MCP Server

Octobrain

Persistent memory for AI assistants — store insights, decisions, and knowledge that survives across conversations.

MCP Registry: mcp-name: io.github.Muvon/octobrain

Table of Contents

• Why Octobrain?
• Quick Start
• Installation
• Usage
• Features
• Benchmarks
• Configuration
• Memory Types
• MCP Integration
• Storage Locations
• Contributing
• License

Octobrain gives your AI assistant a long-term memory. Store code insights, architecture decisions, bug fixes, and knowledge — then retrieve them with semantic search in future sessions. Works as a CLI tool or as an MCP server for integration with Claude Desktop and other AI tools.

Why Octobrain?

AI assistants start every conversation with zero context. You explain your project, your preferences, your decisions — every single time. Octobrain breaks that cycle:

• Persistent memory — Insights survive across sessions, not just within them
• Semantic search — Find memories by meaning, not exact keywords
• Auto-linking — Related memories connect automatically (Zettelkasten-style)
• Knowledge indexing — Ingest docs, articles, and files for retrieval
• MCP integration — Works with Claude Desktop and other MCP-compatible tools

Quick Start

# Install from crates.io
cargo install octobrain

# Store your first memory
octobrain memory memorize --title "API Design Pattern" \
  --content "Use REST for CRUD, GraphQL for complex queries" \
  --memory-type architecture --tags "api,design"

# Search memories
octobrain memory remember "how should I design APIs"

# Start MCP server for Claude Desktop integration
octobrain mcp

Installation

From crates.io (Recommended)

cargo install octobrain

From Source

# Clone and build
git clone https://github.com/muvon/octobrain.git
cd octobrain
cargo build --release

# Binary location
./target/release/octobrain --help

Docker

docker pull ghcr.io/muvon/octobrain:latest

Homebrew (macOS/Linux)

brew install muvon/tap/octobrain

Feature Flags

Octobrain supports multiple embedding providers:

Flag	Description	API Key Required
fastembed	Local embeddings via FastEmbed	No
huggingface	Local embeddings via HuggingFace	No
(default)	Both fastembed + huggingface	No
(no features)	API-based: Voyage, OpenAI, Google, Jina	Yes

# Build with local embeddings (default, no API keys needed)
cargo build --release

# Build with API-based embeddings only
cargo build --no-default-features --release

For API-based embeddings, set the appropriate environment variable:

• VOYAGE_API_KEY for Voyage AI
• OPENAI_API_KEY for OpenAI
• GOOGLE_API_KEY for Google
• JINA_API_KEY for Jina

Usage

Memory Management

Store and retrieve insights, decisions, and context:

All memory subcommands accept global flags:

• --scope <string> — Override project scope (default: auto-detected from Git remote)
• --role <string> — Filter by role (e.g. "developer", "reviewer")

# Store a memory
octobrain memory memorize --title "API Design" \
  --content "Use REST for CRUD, GraphQL for complex queries" \
  --memory-type architecture --tags "api,design"

# Search memories (semantic search)
octobrain memory remember "api design patterns"

# Multi-query search for broader coverage
octobrain memory remember "authentication" "security" "jwt"

# Get a memory by ID
octobrain memory get <id>

# Get recent memories
octobrain memory recent --limit 20

# Filter by type
octobrain memory by-type architecture --limit 10

# Filter by tags
octobrain memory by-tags "api,security"

# Find memories related to files
octobrain memory for-files "src/main.rs,src/lib.rs"

# Update a memory
octobrain memory update <id> --title "New Title" --add-tags "new-tag"

# Delete a memory
octobrain memory forget --memory-id <id>

# Get memories relevant to the current Git commit
octobrain memory current-commit

# Show memory statistics
octobrain memory stats

# Clean up old/low-importance memories
octobrain memory cleanup

# Compact tables and refresh indices (run when searches slow down)
octobrain memory maintenance

# ⚠️ Delete ALL memory data
octobrain memory clear-all --yes

Memory Consolidation

Close a goal and fold all its contributing memories into a consolidated summary:

# Consolidate a goal (Achieves-link sources → Consolidated state, importance dampened)
octobrain memory consolidate <goal-id> --summary "Final summary"

# Sleep consolidation: auto-cluster recent similar memories
octobrain memory sleep-consolidate --threshold 0.85 --min-size 3

Memory Relationships

Connect related memories for context-rich retrieval:

# Create a relationship between memories
octobrain memory relate <source-id> <target-id> \
  --relationship-type "depends_on" \
  --description "Source requires target to function"

# View relationships for a memory
octobrain memory relationships <memory-id>

# Find related memories through relationships
octobrain memory related <memory-id>

# Auto-link similar memories (Zettelkasten-style)
octobrain memory auto-link <memory-id>

# Explore memory graph
octobrain memory graph <memory-id> --depth 2

Knowledge Base

Index and search web content, docs, and files:

# Index a URL
octobrain knowledge index https://docs.rs/tokio/latest/tokio/

# Index a local file (.txt, .md, .pdf, .docx, .html)
octobrain knowledge index ./docs/architecture.md

# Search knowledge base
octobrain knowledge search "how to handle async tasks"

# Search within a specific source (auto-indexes if outdated)
octobrain knowledge search "spawn blocking" --source https://docs.rs/tokio/

# Read full content of a URL or local file
octobrain knowledge read https://docs.rs/tokio/latest/tokio/

# Search indexed content by regex pattern
octobrain knowledge match "spawn_blocking|block_in_place"

# Store raw text content
octobrain knowledge store "meeting-notes" --content "Discussion points..."

# List indexed sources
octobrain knowledge list --limit 20

# Show statistics
octobrain knowledge stats

# Delete a source
octobrain knowledge delete https://example.com/docs

# Delete stored content by key
octobrain knowledge delete-stored "meeting-notes"

Knowledge Boxes

Import and sync git-backed knowledge bundles scoped to your projects:

# Import a remote git repo as a knowledge box
octobrain box import https://github.com/org/docs-repo.git

# Import at global scope (visible in every project)
octobrain box import https://github.com/org/docs-repo.git --global

# Pull and re-index all subscribed boxes + local .box/ directories
octobrain box sync

# List subscribed boxes
octobrain box list

# Remove a box
octobrain box remove github.com/org/docs-repo

MCP Server

Run as an MCP server for integration with Claude Desktop and other AI tools:

# Start with stdio transport (for Claude Desktop)
octobrain mcp

# Start with HTTP transport (for web-based tools)
octobrain mcp --bind 0.0.0.0:12345

Available MCP Tools:

Tool	Description
memorize	Store memories with metadata; optional related_to for inline relationships
remember	Semantic search with filters; returns 1-hop graph neighbors
forget	Delete memories (requires confirmation)
knowledge	Unified tool: search, store, delete, read, match via command field
See MCP Integration for Claude Desktop setup.

Features

• Semantic Search — Find memories by meaning using vector embeddings, not exact keyword matches
• Hybrid Search — Combines BM25 full-text search with vector similarity for better results
• Reranking Support — Pluggable cross-encoder reranking stage (provider-dependent)
• Auto-Linking — Automatically connects semantically similar memories (Zettelkasten-style)
• Temporal Decay — Ebbinghaus forgetting curve for importance management
• Knowledge Indexing — Ingest URLs, PDFs, docs for retrieval
• Project Scoping — Isolate memories per Git project or share across projects
• Role Filtering — Tag memories by role (developer, reviewer, etc.)
• Query Expansion (HyDE-lite) — Pseudo-relevance feedback for +10-30% recall on long-tail queries
• MCP Protocol — Full MCP 2026-07-28 compliance for AI tool integration

Benchmarks

Retrieval quality of octobrain's knowledge system on standard BEIR datasets — nDCG@10, fully local, no LLM judge, using the harness's default embedder bge-small-en-v1.5 (384-dim, 33M params). Each corpus passage is indexed through octobrain's real retrieval path and scored against the official qrels (metrics reproduce pytrec_eval).

Dataset	octobrain vector	octobrain hybrid	BM25¹	bge-small-en-v1.5²
SciFact (5.2K docs, 300 q)	0.722	0.742	0.665	0.713
NFCorpus (3.6K docs, 323 q)	0.341	0.363	0.325	0.343

• vector = dense-only retrieval; reproduces the embedder's published BEIR numbers (validates the harness).
• hybrid = BM25 + vector fused with Reciprocal Rank Fusion (k=60) — octobrain's default. Adds +2 nDCG@10 over the bare embedding and beats classic BM25 on both datasets.

¹ Canonical BM25 from the BEIR paper (Anserini/Lucene, k1=0.9 b=0.4). ² From the bge-small-en-v1.5 model card (MTEB).

Scope: this measures the ranking layer (embedding + BM25 fusion + reranking). BEIR passages are pre-chunked, so octobrain's chunking strategy is not exercised here.

⚠️ Reranker status: The cross-encoder reranker is currently a no-op with fastembed models (byte-identical results to hybrid-only). Rerank is excluded from the numbers above pending investigation. See benches/README.md for details.

Reproduce (downloads the datasets, builds a release binary, runs fully offline):

cd benches && bash scripts/run_retrieval.sh

Configuration

Configuration is stored in ~/.local/share/octobrain/config.toml. The template written on first run defines every option with sensible defaults, but loading is strict — a field you delete by hand causes startup to fail rather than falling back to a default.

Override the config location with OCTOBRAIN_CONFIG_PATH=/path/to/config.toml.

Key Settings

Section	Option	Default	Description
[embedding]	model	fastembed:Qdrant/all-MiniLM-L6-v2-onnx	Embedding model (provider:model format). Default is a local fastembed model — no API key, runs on CPU.
[search]	similarity_threshold	0.3	Minimum relevance (0.0-1.0)
[search.hybrid]	enabled	true	Enable BM25 + vector fusion
[search.reranker]	enabled	true	Enable cross-encoder reranking
[search.hyde]	enabled	true	Pseudo-relevance feedback query expansion
[memory]	max_memories	10000	Maximum stored memories
[memory]	auto_linking_enabled	true	Auto-connect similar memories
[knowledge]	chunk_size	1200	Characters per chunk
[embedding]	batch_size	32	Texts embedded per batch
[embedding]	timeout_secs	30	Embedding call timeout (0 = no timeout)
[search]	max_results	50	Hard ceiling on results from any search
[search.reranker]	model	fastembed:jina-reranker-v2-base-multilingual	Cross-encoder model
[search.reranker]	top_k_candidates	50	Candidates retrieved before reranking
[search.hyde]	top_k	3	Neighbors averaged for the centroid
[search.hyde]	alpha	0.5	Blend weight on the original query embedding
[memory]	max_search_results	50	Default page size when no limit is given
[memory]	sleep_consolidation_enabled	true	Lazy sleep consolidation on manager init
[memory]	sleep_consolidation_interval_hours	24	Hours between sleep-consolidation passes
[knowledge]	outdating_days	15	Days before indexed content is reindexed on search
[knowledge]	max_results	5	Results returned from knowledge search

Embedding Providers

[embedding]
# Local models (no API key, runs on CPU, model auto-downloaded on first use)
model = "fastembed:Qdrant/all-MiniLM-L6-v2-onnx"                       # Default: 384-dim quantized, ~125MB resident
model = "fastembed:nomic-ai/nomic-embed-text-v1.5"                      # 768-dim, 8192-ctx, ~550MB weights, highest quality
model = "fastembed:BAAI/bge-small-en-v1.5"                              # 384-dim, ~62 MTEB, fast + good quality
model = "fastembed:sentence-transformers/all-MiniLM-L6-v2-quantized"  # Smallest (~22MB), fastest
model = "fastembed:BAAI/bge-base-en-v1.5"                              # Larger (~440MB), higher quality
model = "fastembed:intfloat/multilingual-e5-small"                     # Multilingual

# Cloud providers (require API keys, generally higher quality)
model = "voyage:voyage-3.5-lite"          # VOYAGE_API_KEY
model = "openai:text-embedding-3-small"   # OPENAI_API_KEY
model = "google:text-embedding-004"       # GOOGLE_API_KEY
model = "jina:jina-embeddings-v3"         # JINA_API_KEY

Full Configuration

See config-templates/default.toml for all available options with documentation.

Memory Types

Organize memories by category for better filtering:

Type	Use For
code	Code patterns, solutions, implementations
architecture	System design, decisions, patterns
bug_fix	Bug fixes, troubleshooting, solutions
feature	Feature specs, implementations
documentation	Docs, explanations, knowledge
user_preference	Settings, preferences, workflows
decision	Project decisions, trade-offs
learning	Tutorials, notes, education
configuration	Setup, config, deployment
testing	Test strategies, QA insights
performance	Optimizations, benchmarks
security	Vulnerabilities, fixes, considerations
validation	Idea/product validation, hypothesis testing
research	Technical/market research, analysis
workflow	SOPs, playbooks, process descriptions
requirement	Business requirements, specs, constraints
design	UI/UX decisions, wireframes, system design
integration	API integrations, third-party services
communication	Stakeholder updates, team decisions
process	Deployment procedures, runbooks, operations
insight	General insights, tips
goal	Task/intent anchors for consolidation workflow

MCP Integration

Claude Desktop Setup

Add to your Claude Desktop config (~/Library/Application Support/Claude/claude_desktop_config.json on macOS):

{
  "mcpServers": {
    "octobrain": {
      "command": "/path/to/octobrain",
      "args": ["mcp"]
    }
  }
}

Restart Claude Desktop. Octobrain tools will be available in your conversations.

HTTP Transport

For web-based integrations:

octobrain mcp --bind 0.0.0.0:12345

The server exposes endpoints at /mcp for MCP protocol communication.

Storage Locations

Data is stored in platform-specific directories:

Platform	Location
macOS	~/.local/share/octobrain/
Linux	~/.local/share/octobrain/ or $XDG_DATA_HOME/octobrain/
Windows	%APPDATA%\octobrain\

Project-specific memories are isolated by normalized Git remote URL (e.g., github.com/org/repo).

Contributing

Contributions are welcome! Please:

1. Fork the repository
2. Create a feature branch (git checkout -b feature/amazing-feature)
3. Run cargo fmt --all to format code
4. Run cargo clippy and fix all warnings
5. Run cargo test --no-default-features
6. Submit a pull request

Development Setup

# Clone and build
git clone https://github.com/muvon/octobrain.git
cd octobrain
cargo build --no-default-features

# Run tests
cargo test --no-default-features

# Run clippy
cargo clippy --no-default-features

License

Apache-2.0 — see LICENSE for details.

Credits

Developed by Muvon Un Limited.