r0b0tlabbra1n
r0b0tlab/llm-wiki_obsidian_hermes_r0b0tlabbra1n
Filesystem-first Markdown and Obsidian memory vault for Hermes with SQLite FTS5 search and secret scanning
r0b0tlabbra1n is a local, auditable memory system for Hermes Agent and other CLI agents, built on a Markdown and Obsidian vault. It adds SQLite search, linting, session ingest and a Hermes skill pack.
What r0b0tlabbra1n does
r0b0tlabbra1n is a local, auditable memory system for Hermes and other CLI agents. It keeps durable knowledge in human-editable Markdown, uses Obsidian wikilinks for navigation, and treats the Markdown as the source of truth. Everything else, including the SQLite FTS5 search index, the link graph and the source-hash manifest, can be rebuilt from the vault at any time.
A brain command line handles vault initialization, linting for frontmatter, wikilinks and secrets, BM25 search with tier, confidence and recency weighting, context packets with a size budget, drift checks, graph and backlinks output, and a gold-query evaluation harness reporting top-1, top-3 and MRR. It can ingest Hermes sessions read-only from ~/.hermes, generate promotion candidates, and ships a Hermes skill pack, cron heartbeat and audit scripts, and a minimal MCP and JSON facade. A secret scanner blocks credentials from entering the vault write path.
Key features
- Markdown vault with Obsidian wikilinks as the source of truth
- SQLite FTS5 search with BM25 ranking and tier-aware context packets
- Linting for frontmatter, wikilinks and secrets
- Read-only ingest of Hermes sessions plus promotion candidate generation
- Link graph, backlinks and source-hash drift checks
- Gold-query retrieval eval harness with top-1, top-3 and MRR
When to use it
- Give Hermes a human-readable long-term memory that you can edit and diff
- Search past Hermes sessions and project notes with ranked local retrieval
- Keep API keys and tokens out of agent memory with write-path secret scanning
Who it is for: Hermes Agent users running local LLM setups who want an auditable Markdown memory instead of opaque stores.
How it fits with Hermes Agent
Built for Hermes and other CLI agents, with a Hermes skill pack, cron scripts and a session ingest that reads from the Hermes home directory.
How to install r0b0tlabbra1n
These commands are copied from the project's README. Check the repository for the latest steps before you run them.
git clone https://github.com/r0b0tlab/llm-wiki_obsidian_hermes_r0b0tlabbra1n ~/brain-code
cd ~/brain-code
python3 -m venv .venv
.venv/bin/pip install -e ".[dev]"Note: Optional vector embeddings and an advanced temporal graph database are listed as roadmap items and are not implemented yet.
FAQ
What is r0b0tlabbra1n?
r0b0tlabbra1n is a filesystem-first memory system that combines an LLM wiki, an Obsidian vault and Hermes Agent. Markdown is the source of truth, and the search and graph indexes can be rebuilt from it.
Does r0b0tlabbra1n work with Hermes Agent?
Yes. It ingests Hermes sessions read-only from the Hermes home directory and includes a Hermes skill pack, heartbeat and audit scripts, and a minimal MCP and JSON facade.
Is r0b0tlabbra1n free and open source?
Yes. The repository is released under the MIT license.
Similar memory for Hermes Agent
All memoryGuide to running Honcho and GBrain together with Hermes Agent for relationship and source memory
smarzola Hermes Local MemoryLocal-first SQLite memory provider for Hermes Agent with peers, cards, search and consolidation
ncz-os MNEMOSMemory operating system for AI agents with MCP, OpenAI-compatible and REST interfaces
M4F-S GomaaLocal-first hierarchical memory engine with Obsidian vault storage, hybrid search and an MCP server
TJCurnutte Memory HiveFile-backed shared memory for AI agent teams with private silos, a shared hive and a curator workflow
gymaira1990-jpg Mnemosyne OSSelf-hosted memory palace for AI agents with PostgreSQL, pgvector and a Hermes memory provider
Related guides: SOUL.md for Hermes Agent: what it is and how to write one · Run multiple Hermes agents with profiles