Memware — Cross-Agent Long-Term Memory
Shared long-term memory that carries your habits and open tasks across AI coding agents
The server answered 3 hours ago with a sign-in request for an API key (HTTP 401), so we could not list its tools.
- Pricing
- Freemium
- Platforms
- Web, iOS, Android
- Runs
- Remotely
- Access
- Needs sign-in (API key)
- Version
- 0.2.1
- Listing updated
- today
Similar MCP servers
About Memware — Cross-Agent Long-Term Memory
Memware — Cross-Agent Long-Term Memory is a freemium MCP server for knowledge and memory. It runs as a hosted remote server, so there is nothing to install locally.
The README says Memware keeps a single vault of preferences, facts and unfinished engineering threads that Claude Code, Codex, Cursor, Antigravity and the Claude and ChatGPT mobile apps can all read. Instead of relying on the model to decide when to save something, it uses Stop hooks to capture each completed turn in the background, so context persists without an explicit save command.
The product runs in two modes. Memware Cloud connects through a remote MCP endpoint with a bearer token and needs no LLM keys from the user. Memware Local runs as a stdio server backed by a SQLite database on the user's machine, using an OpenAI-compatible key the user supplies. Both expose the same 14 MCP tools, including budgeted context recall, hybrid search, and a memory_resume handoff briefing for moving a task from one agent to another.
It suits developers who switch between several AI coding tools and want their project context to follow them, and people who want to inspect, edit or fully erase what an assistant has stored. Vaults can be separated into work and personal scopes on the paid tiers.
Install
Run in a terminal
claude mcp add --transport http memware https://mcp.aiware.life/mcpAdd to Settings → Connectors → Add custom connector
https://mcp.aiware.life/mcpAdd to ~/.cursor/mcp.json
{
"mcpServers": {
"memware": {
"url": "https://mcp.aiware.life/mcp"
}
}
}Add to .vscode/mcp.json
{
"servers": {
"memware": {
"type": "http",
"url": "https://mcp.aiware.life/mcp"
}
}
}Add to ~/.codeium/windsurf/mcp_config.json
{
"mcpServers": {
"memware": {
"serverUrl": "https://mcp.aiware.life/mcp"
}
}
}Add to Grok Bot desktop app → Connect apps → Add a custom connector
https://mcp.aiware.life/mcpEnter a Name (for example "Memware — Cross-Agent Long-Term Memory"), choose Type: Remote HTTPS, paste this as the Server URL and choose Add plugin. When we checked, this server asked for an API key. Add it under HTTP headers, using the header name from the server's docs.
Add to your client's MCP settings (most use this shape)
{
"mcpServers": {
"memware": {
"type": "http",
"url": "https://mcp.aiware.life/mcp"
}
}
}Using several servers? Build one config file with the MCP config generator.
Tools this server provides (14)
Taken from the project documentation; the running server may differ.
- memory_status
- Inspect vault status, active tier, quota usage, and configured model origins.
- memory_warmup
- Warm up and initialize the bound user profile idempotently.
- memory_get_context
- Assemble budgeted context from profile, active threads and hybrid semantic recall for a query.
- memory_profile
- Read structured user profile facts, preferences, and standing constraints.
- memory_remember
- Explicitly set or overwrite a durable user preference.
- memory_threads
- List active unfinished work threads with last-agent provenance and next steps.
- memory_process
- Distill and deduplicate durable facts and thread state from a conversation turn.
- memory_search
- Hybrid dense-vector and lexical search across stored memories.
- memory_list
- Paginate stored memory records newest-first.
- memory_update
- Edit an existing memory fact, category, or importance score in place.
- memory_delete
- Soft-delete a specific memory record and remove its vectors.
- memory_resume
- Generate a ready-to-continue briefing for cross-agent task handoff.
- memory_archive
- Archive stale low-activity memory clusters.
- memory_reset
- Physically wipe all memories, vectors and threads in the vault.
What it does
- Automatic turn capture
- Stop hooks save habits and unfinished tasks after each completed turn in supported agents.
- Cross-agent task handoff
- The memory_resume tool produces a briefing so another agent can pick up an open thread.
- Cloud or fully offline mode
- Use the hosted vault over remote MCP, or run a local SQLite store with no cloud dependency.
- Hybrid memory search
- Combines vector and lexical search to recall profile facts and active tasks within a call.
- Inspectable and erasable memory
- Edit, delete or reset stored memories, with memory_reset wiping records, vectors and audit logs.
- Separate vaults
- Scoped vault tokens keep work and personal context apart on the paid plans.
Good for
- Continue an unfinished coding task in a different agent without re-explaining the project
- Keep personal coding preferences consistent across Claude Code, Codex and Cursor
- Check on open work threads from the official Claude or ChatGPT mobile app
- Keep work and personal AI context in separate, non-shared vaults
- Run an assistant memory entirely on a local machine with its own SQLite database
Pricing
Freemium, from $12 a month
Free tier with 200 active memories; Pro from $12/mo ($99/yr) and Max at $29/mo per the README; Local Lifetime at $49 early bird or $69 standard.
Prices change. Check aiware.life before you buy.
Memware — Cross-Agent Long-Term Memory FAQ
How do I connect Memware to Claude Code?
The README gives a command using claude mcp add with the HTTP transport to https://mcp.aiware.life/mcp, passing a mw_live token in an Authorization header.
Is there a free plan?
The pricing page lists a free tier with up to 200 active memories and a local CLI free tier that uses your own key. Cloud free-tier details are listed as up to 200 active memories and 150 cloud auto-extractions per month.
Can I erase everything Memware stores?
Yes. The site and README describe memory_reset, which purges the tenant SQLite database, vector index, relation projections and audit logs in both cloud and local modes.
Does the local version need an internet connection?
The site says the local binary verifies MW1 license keys offline with Ed25519 signatures and does not phone home. Local mode uses your own OpenAI-compatible key for model calls.
Is Memware — Cross-Agent Long-Term Memory free?
Memware — Cross-Agent Long-Term Memory has a free plan, with paid plans for more usage or features. Paid plans start at $12 a month.
How do I add Memware — Cross-Agent Long-Term Memory to Claude Code?
Run this command in your terminal: claude mcp add --transport http memware https://mcp.aiware.life/mcp
What are the best alternatives to Memware — Cross-Agent Long-Term Memory?
The closest alternatives to Memware — Cross-Agent Long-Term Memory are memry, Memmo, Memtomem, Memnest and Mnemosyne OS.
Where this comes from
- Description: written from the project's README, the project's website and its MCP Registry entry.
- Status: checked automatically by AIPoweredBot on 19 November 2026.
- Install steps: from the official MCP Registry entry.