CogniCore
Shared memory for Claude, ChatGPT and coding agents so new chats start informed
- Pricing
- Open source
- Platforms
- CLI, Web, self-hosted
- Runs
- Remotely
- Version
- 0.10.3
- Listing updated
- 2 days ago
About CogniCore
CogniCore is an open-source MCP server for knowledge and memory. It runs as a hosted remote server, so there is nothing to install locally.
Every new chat starts blank, so decisions and context get re-explained. CogniCore saves a short brief of the goal, where things stand, open tasks, decisions with their reasons and exact values like ticket IDs, versions and paths, then loads it at the next start. In Claude Code and Codex hooks do this automatically when a session ends; other MCP clients can call cognicore_save_session and cognicore_resume_session.
Memory is local by default, kept in SQLite under ~/.cognicore/mcp, and secrets are redacted before saving. The README says it never reads tool output, file contents or the model's reasoning. Running cognicore login syncs only the redacted handoff to a hosted account so claude.ai, Claude Desktop, mobile and ChatGPT can share it. The README includes a benchmark against claude-mem, reporting about 450 tokens for the injected brief and no model calls to save a session.
Underneath is a Python library for agent memory, reflection and safety that needs no vector database or embedding server. The site says the code is open source and that data is not sold or used for training.
Install
Run in a terminal
claude mcp add --transport http cognicore-env https://mcp.cognicoree.tech/mcp/streamAdd to Settings → Connectors → Add custom connector
https://mcp.cognicoree.tech/mcp/streamAdd to ~/.cursor/mcp.json
{
"mcpServers": {
"cognicore-env": {
"url": "https://mcp.cognicoree.tech/mcp/stream"
}
}
}Add to .vscode/mcp.json
{
"servers": {
"cognicore-env": {
"type": "http",
"url": "https://mcp.cognicoree.tech/mcp/stream"
}
}
}Add to ~/.codeium/windsurf/mcp_config.json
{
"mcpServers": {
"cognicore-env": {
"serverUrl": "https://mcp.cognicoree.tech/mcp/stream"
}
}
}Add to your client's MCP settings (most use this shape)
{
"mcpServers": {
"cognicore-env": {
"type": "http",
"url": "https://mcp.cognicoree.tech/mcp/stream"
}
}
}Tools this server provides (2)
Taken from the project documentation; the running server may differ.
- cognicore_save_session
- Saves a session handoff for later resumption.
- cognicore_resume_session
- Loads the latest saved handoff for a project.
What it does
- Cross-agent handoff
- Work saved in Claude Code resumes in Codex and the other way round.
- Automatic save
- Hooks save a session when it ends, with no model calls.
- Compact brief
- Injects about 450 tokens covering goal, status, tasks and decisions.
- Local by default
- Stores memory in SQLite on your machine and redacts secrets.
- Optional hosted sync
- Shares the redacted handoff with claude.ai, Desktop, mobile and ChatGPT.
- Inspectable
- A read-only viewer lists every saved session.
Good for
- Continue a Claude Code task in Codex without re-explaining
- Keep preferences and decisions across Claude and ChatGPT chats
- Resume a project the next morning from a short brief
- Add persistent memory to a custom Python agent
Pricing
Open source
Free and open source; hosted memory is optional
Prices change. Check cognicoree.tech before you buy.
CogniCore FAQ
How do I install it?
Run pipx install "cognicore-env[mcp]" then cognicore install, and cognicore login for hosted sync.
Where is memory stored?
Locally in ~/.cognicore/mcp unless you turn on sync.
Is it free?
The site describes it as free and open source.
Does it need a vector database?
No, the library needs none.
What does it send when syncing?
Only the redacted handoff, never the transcript.
Is CogniCore free?
Yes. CogniCore is open source, so you can use and self-host it for free.
How do I add CogniCore to Claude Code?
Run this command in your terminal: claude mcp add --transport http cognicore-env https://mcp.cognicoree.tech/mcp/stream
What are the best alternatives to CogniCore?
The closest alternatives to CogniCore are Codebase Memory, XERJ, Claude Code Ultimate Guide, Nautilus Compass Memory and note1.
Alternatives to CogniCore

Codebase Memory
MCP serverLocal MCP server that builds a code knowledge graph for AI coding agents
Runs locally- Open source
- 48.7K GitHub stars
- Knowledge and memory

XERJ
MCP serverLocal search engine for AI agents with cited answers over your code and docs
Runs locally- Open source
- 19.2K GitHub stars
- Knowledge and memory

Claude Code Ultimate Guide
MCP serverMCP server for searching the Claude Code Ultimate Guide from any MCP client
Runs locally- Open source
- 6.5K GitHub stars
- AI models and LLM tooling

Nautilus Compass Memory
MCP serverLocal-first long-term memory layer for AI agents, served over MCP
Not checked yet- 5 tools
- Remote
- Open source
- 1.3K GitHub stars
- Knowledge and memory

note1
MCP serverMCP server giving AI assistants search and summaries across your note1 meetings
- 4 tools
- Remote
- Free trial
- 1 GitHub star
- Productivity
MedXpert 图书馆连接器(通用版)
MCP serverLocal MCP server that searches and answers questions over your own Markdown wiki
Runs locally- 4 tools
- Open source
- Knowledge and memory