Connect your AI in minutes
MemScope speaks MCP for assistants and agents, and a plain REST API for internal apps.
Connect Claude Code
claude mcp add --transport http --scope user memscope https://mcp.memscope.ai/mcp
Connect ChatGPT
MemScope is not in the ChatGPT app directory yet, so you add it as a custom connector (turn on Developer mode first). Connecting is an OAuth handshake: find MemScope, sign in, choose the organization, approve the permissions MemScope asks for. Your ChatGPT session then acts as you — never as a privileged robot.
- 1In ChatGPT, open Settings → Connectors and add MemScope.
- 2Sign in and pick the organization to connect.
- 3Review the permissions MemScope asks for on the consent screen and approve.
- 4Ask ChatGPT to save something to MemScope, then ask it back. It calls save_memory and search_memories.
Need the ChatGPT app submission details? See CHATGPT_MCP_SETUP.md in the repository.
Connect other MCP clients
Any MCP-capable client (Claude, Cursor, a custom agent) can connect to the remote MCP endpoint over streamable HTTP. Auth is OAuth 2.1 with PKCE, dynamic client registration and a consent screen.
https://mcp.memscope.ai/mcp
Discovery is automatic via /.well-known/oauth-protected-resource. Clients register themselves, then the user approves access.
Exact commands and config for Claude Code, Cursor, VS Code, Codex, Gemini CLI and more — written so an AI agent can follow them — are on MCP setup for AI agents (plain text: /llms.txt).
Save your first memory
Once your AI is connected, you don't call the tools yourself: ask in plain language and it calls them for you.
- 1Ask: “Remember that Project Falcon switched to Supplier B.” The AI calls save_memory.
- 2In a new chat, ask: “Which supplier does Project Falcon use?” The AI calls search_memories and answers from the memory.
- 3Open Memories in MemScope to see it, with its audience and sensitivity.
- 4Ask: “Who can see that memory?” The AI calls explain_memory and names the rule.
MCP tools reference
save_memorycontent, scope?, team?, teams?, project?, importance?, idempotency_key?, kind?, valid_until?, supersedes?, agent?, run_id?Store an explicit memory. Classified asynchronously; returns immediately. (Old name: remember.)
search_memoriesquery, project?, team?, topics?, teams?, limit?, detail?, max_chars?, kind?, include_expired?, include_superseded?, run_id?Search memory by meaning. Only memories the authenticated human may read are returned. (Old name: recall.)
grep_memoriespattern, ignore_case?, teams?, projects?, topics?, kind?, include_expired?, include_superseded?, limit?, cursor?Find exact text (names, IDs, error strings), like grep: matching lines in context, newest first, with counts per team, project and tag. Only memories the person may read are searched.
list_teamskind?List the teams and projects the signed-in person's organization exposes — the valid values for `search_memories`' team/project filters and `save_memory`'s filing targets. kind is teams, projects or all (default all). Never lists another organization's groups.
list_topicsList the organization's sensitive topics (HR, Legal, Finance and any the org added) — good `search_memories` query terms for guarded subjects. Read-only; only the caller's organization's topics.
explain_memorymemory_idWhy a memory is or is not accessible, including the governing rule.
correct_memorymemory_id, new_content, reason, kind?, valid_until?, agent?, run_id?Create a new immutable version of a memory with a corrected body.
forget_memorymemory_id, reason, confirm_human?, agent?, run_id?Forget a memory, only when the person asks (needs their confirmation). Authorized users can restore it for 30 days; then it is erased for good.
list_memoriesfilter?, limit?, cursor?, kind?, agent?, run_id?Browse memories the person may read (filter: mine, needs_review, team, project, forgotten).
get_memorymemory_idFull detail of one memory, including its version history.
change_scopememory_id, scope, reason, team?, teams?, project?, confirm_widen?, confirm_human?, agent?, run_id?Change who may read a memory. A change that could give more people access needs the person’s confirmation.
change_sensitivitymemory_id, sensitivity, reason, confirm_widen?, confirm_human?, agent?, run_id?Change a memory’s sensitivity. Lowering it needs the person’s confirmation.
review_queuelimit?Flagged memories waiting for the person to review.
review_memorymemory_id, action, reason, scope?, team?, project?, sensitivity?, confirm_widen?, confirm_human?, agent?, run_id?Resolve a review: approve, reroute to a team or project, or tighten.
supersede_memorymemory_id, replaced_by, undo?, agent?, run_id?Mark a memory as replaced by a newer one; it leaves default recall, history kept.
restore_memorymemory_id, reason, agent?, run_id?Bring back a forgotten memory.
REST API
Internal applications use the same memory service over REST. Browser clients authenticate with an HttpOnly session cookie; programmatic clients may use an OAuth bearer token.
Interactive API reference
https://api.memscope.ai/docscurl -X POST https://api.memscope.ai/organizations/acme/memories \
-H 'Content-Type: application/json' \
-H 'X-Requested-With: memscope' \
-b cookies.txt \
-d '{"content":"Project Falcon switched to Supplier B.","scope":"project"}'Ready to connect?
Create an organization and generate your first connection.
Start free