Install Knos in Claude Code
Remote URL
claude mcp add --transport http knos https://knos.onrender.com/mcp{
"mcpServers": {
"knos": {
"url": "https://knos.onrender.com/mcp"
}
}
}PyPI (uv)
claude mcp add knos -- uvx knos{
"mcpServers": {
"knos": {
"command": "uvx",
"args": [
"knos"
]
}
}
}Frequently asked questions
What does the Knos MCP server do?
Local memory shared by all coding agents on a machine, built from sessions, commits, CLAUDE.md/AGENTS.md and ADRs, with claims that lock in-progress work.
Is the Knos MCP server free?
Yes, most likely. Open source and no paid account or API key is mentioned in the listing. Check the provider for current pricing.
Does Knos run locally or remotely?
Local + Remote. Can run locally or be used as a hosted remote server.
How do I add Knos to Claude Code?
Run: claude mcp add --transport http knos https://knos.onrender.com/mcp
More knowledge, memory & rag MCP servers
View allLinkLore
linklore/linklore-mcp
Structured memory for coding agents: typed lore and doc entries with status and links that agents read and write.
JidoSeal
Jidoseal/jidoseal-mcp
Checks a folder of Markdown docs for missing Open Knowledge Format (OKF) frontmatter fields, locally; writes only .jidoseal/ in that folder. MCP server Apache-2.0, scan engine proprietary; no network use during a scan is a vendor claim.
LocalKeep
zuneracodes/localkeep
Local-first chronological memory: write dated, editable memory banks in a local UI, and Claude recalls them on demand via SQLite FTS5.
Echocache
kskurtveit/echocache
Cache expensive LLM results with HTTP-style freshness (TTL, stale-while-revalidate) and semantic recall of related past answers.
md-vision
japlete/md-vision-mcp
Read Markdown with inlined images and heading indexes for agentic RAG over local docs or allowed URLs.
Compound Memory
chinwe/compound-memory
Local-first shared memory for multiple AI agents that compounds in value: feedback-driven confidence, cross-agent validation, neighbor recall and distillation; Markdown files with auto git history, 5 MCP tools + CLI.