Wing
A project or context namespace — one isolated workspace per team.
Open-source · MCP · multi-tenant
Long-term memory for your AI agents.
AI Agent Memory is an open-source, multi-tenant agent memory server. Your agents connect over MCP, file verbatim memories, and recall them with hybrid semantic search — so every session builds on the last instead of starting from zero.
Free Personal plan · 10,000 requests / month · no card required
36 / 37
MCP tools shipped
3-way
hybrid recall: vector · BM25 · closet
per-team
isolated vector store
$0
to start — 10k requests / month
The idea
What is AI agent memory?
An LLM forgets everything the moment its context window closes. AI agent memory fixes that: it is durable, long-term storage that an AI agent reads from and writes to across sessions — the decisions it made, the facts it learned, the threads it left open.
AI Agent Memory serves that store as a remote MCP server. Agents file verbatim memories — never lossy summaries — and recall the right ones on demand with hybrid semantic search. The next run picks up exactly where the last one stopped.
The data model
The memory palace is the schema
Seven primitives compose every memory — borrowed from how humans file what they want to remember.
Room
An aspect within a wing, like backend or decisions.
Drawer
One verbatim memory chunk plus rich metadata. Never summarised.
Closet
A topic and quote pointer index that boosts ranking — never a gate.
Hallway
A within-wing link between entities that co-occur in drawers.
Tunnel
A cross-wing link — authored, or auto-derived from a shared topic.
Knowledge graph
Temporal subject→predicate→object facts with validity windows.
How it works
One endpoint, fully isolated
Stateless Streamable-HTTP MCP: every request re-resolves its tenant from the bearer token, so the service scales out behind a load balancer.
Connect over MCP
Point any MCP client — Claude, your own agent — at POST /mcp with an Authorization: Bearer token.
Resolve the tenant
The token becomes a workspace in exactly one place. Every tool reads that tenant off the context and fails closed without it.
File and recall
Write verbatim drawers that get embedded and indexed, then recall them with hybrid search across the whole team's memory.
Stay isolated
SQLite is the relational source of truth; Qdrant holds per-tenant vectors, rebuildable from it. The transport is stateless, so it scales out.
Capabilities
Everything an agent needs to remember
36 of 37 MCP tools shipped — the write, recall, graph, knowledge-graph and skill families.
Recall
Hybrid semantic search
Vector similarity, BM25 lexical match and a closet boost, fused into one ranking — so agents recall by meaning and by exact term.
Isolation
Memory that can't leak
Every workspace gets its own Qdrant collection, named by a hash of the team id. A missing filter can't cross tenants — the data isn't even colocated.
Skills
Centralised, versioned skills
One shared source of truth for prompts and skills. Agents pull the latest with am_load_skill instead of copy-pasting local files.
Diary
An append-only agent diary
A timestamped journal per agent. Sessions thread across time, so the next run reads what the last one learned.
Knowledge
Temporal knowledge graph
Subject→predicate→object facts with validity windows, queryable as-of any point in time. Know what was true then, not just now.
Mining
Idempotent mining pipeline
am_mine turns raw text into chunked, embedded drawers plus a closet index — keyed by source, so re-running finishes rather than duplicates.
Graph
A navigable memory graph
Hallways link co-occurring entities; tunnels bridge wings. Traverse the graph to surface context a flat search would miss.
Migrate
Bring your mempalace
A read-only exporter streams an existing local mempalace into your workspace over /import — re-embedded server-side, graph rebuilt, fully idempotent.
Quick start
Running in two commands
Go 1.25+. The binary migrates an embedded schema, seeds a demo workspace, and prints a one-time MCP bearer token. Point any MCP client at POST /mcp with that token.
shell
go build -o agentsmemory ./cmd/server
./agentsmemory --addr :8080 --db agentsmemory.db
# prints a one-time MCP bearer token to the logPricing
Start free, scale to a team
One user can own several workspaces across plans, each with its own isolated vector store and its own revocable keys.
Personal
For solo agents and side projects.
$0 forever
- 10,000 requests / month
- Unlimited drawers & diary
- Hybrid search + knowledge graph
- Centralised skills
Enterprise
For teams sharing memory across agents.
$50 / month
- Everything in Personal
- Multiple workspaces & members
- Role-gated shared skills
- Per-team isolated vector store
Migrate
Bring your existing memory palace
Already running the local Python mempalace? A read-only exporter streams every drawer, diary entry, closet, knowledge-graph fact and tunnel into your workspace over /import. The server re-embeds each memory and rebuilds the graph — and the import is idempotent, so a re-run finishes rather than duplicates.
python mempalace_export.py --push \ --server https://aiagentmemory.dev \ --token YOUR_PROJECT_API_KEYFAQ
Questions about agent memory
AI agent memory is persistent, long-term storage that lets an AI agent remember context across sessions — past decisions, facts and learnings — instead of starting cold every run. AI Agent Memory provides it as a remote MCP server: agents file verbatim drawers of memory and recall them later with semantic search.
Give your agents a memory that lasts.
Spin up a free workspace and connect your first agent in minutes.