Local memory infrastructure for AI agents

# Initialize vaults
ctxvault init knowledge
ctxvault init skills --type skill

# Add documents, index, and query
ctxvault index knowledge
ctxvault query knowledge "what decisions were made last week?"

# Inspect skills
ctxvault skills skills
ctxvault skill skills "Weekly Engineering Update"

# Facts that change over time
ctxvault init memory --type episodic
ctxvault write-episode memory "Sarah is the main contact" -e Sarah
ctxvault episodes memory --entity Sarah

# See everything
ctxvault vaults

import requests
from langchain_openai import ChatOpenAI

API = "http://127.0.0.1:8000/ctxvault"

# Write what the agent learns
requests.post(f"{API}/docs/write", json={
    "vault_name": "agent-memory",
    "file_path": "session_monday.md",
    "content": "Q2 budget: cut cloud costs by 15%. Competitor pricing 20% lower."
})

# Days later: query with completely different words
results = requests.post(f"{API}/query", json={
    "vault_name": "agent-memory",
    "query": "financial constraints from last week"
}).json()["results"]

# Ground your LLM in retrieved memory
context = "\n".join(r["text"] for r in results)
answer = ChatOpenAI().invoke(f"Context:\n{context}\n\nQ: What are our cost targets?")

// claude_desktop_config.json
{
  "mcpServers": {
    "ctxvault": {
      "command": "ctxvault-mcp"
    }
  }
}

// The agent handles list_vaults, query, write,
// read_skill, list_skills, write_episode and
// query_episodes autonomously, no glue code.