crystalline
Active·★ 17·AGPL-3.0·Updated 2026-10-21
Crystalline provides AI agents with crystallized intelligence, allowing them to accumulate and retain knowledge across sessions, unlike an LLM's fluid intelligence. It acts as a local search index on plain markdown files, enabling agents to onboard curated knowledge and capture new learnings, evolving from a stranger to a peer.
#AI Agents#Knowledge Base#RAG#Semantic Search#Markdown#Collaboration#Persistent Memory#Dev Tooling
01
Features
01Structured knowledge organization using domains (folders) and engrams (markdown files).
02Hybrid search combining full-text and semantic ranking for efficient knowledge retrieval.
03Fluid web UI for human-agent collaboration, enabling reading, editing, and managing knowledge.
04Supports GitHub-based team collaboration with review workflows for shared knowledge.
05Knowledge graph and temporal filtering to understand relationships and time-bounded facts.
02
Compatibility
macOS
OS
Verified via docs
Linux
OS
Verified via docs
Windows
OS
Verified via docs
Rust
Runtime
Verified via docs
Claude Code
Agent Harness
Verified via docs
Claude Desktop
Agent Harness
Verified via docs
03
Quick start
1
$ brew install jordiboehme/tap/crystalline
2
$ crystalline install claude-code
3
$ mkdir -p ~/knowledge/engineering
4
$ crystalline domain init ~/knowledge/engineering --name engineering
5
$ crystalline domain add engineering ~/knowledge/engineering
04
Use cases
↳Onboard AI agents with curated domain-specific knowledge at the start of each session.
↳Enable AI agents to capture and persist new observations, decisions, and lessons learned from their work.
↳Facilitate team knowledge sharing and collaboration on a continually growing and evolving knowledge base.
↳Reduce redundancy and repeated explanations by providing agents with access to accumulated 'crystallized intelligence'.
↳Efficiently retrieve specific facts or context from a large collection of knowledge units based on task routing.
05
Alternatives
tela★ 84
Open-source, self-hostable markdown team wiki with a built-in MCP server — so AI agents read, write, and search your docs as first-class teammates. Live multiplayer, semantic + full-text search. Go + PostgreSQL + React/Milkdown.
sample-okf-llm-wiki★ 33
Data Wiki turns data into portable knowledge bundles following Open Knowledge Format (OKF) and serves them to AI agents over the Model Context Protocol (MCP).
okf-agent-memory★ 827
Git-native persistent memory for AI coding agents. Implements Google OKF v0.2 with sub-300µs in-memory BM25 search, embedded MCP server, and progressive disclosure. Slashes token bloat by 80% with zero external databases or dependencies. Built in pure Go.
data-olympus★ 30
Governance-grade, OKF-compatible knowledge-base format with a single-writer MCP server and CLI. Pre-release.
Comments
Log in to leave a comment
No comments yet. Be the first!