01
decision
Standardize on idempotent write endpoints
Source: architecture.md
Agent Memory Database in · shared project memory
Agent Memory Database in Seoul
Semantic recall without vendor lock-in: enable pgvector embeddings through Ollama or your own provider, or run pure full-text search. Either way the corpus stays in your PostgreSQL, queryable forever.
JSON files and embedded stores break at team scale. MemHub runs agent memory on PostgreSQL 16 with a pgvector-ready schema: migrations, backups, indexes, and row-level security your DBA already understands.
Create your shared memory
Free Community tier. Self-hosted or hosted. No external embedding API required.
acme/agent-memory-database
synchronizedcontext/main
memories 41
pinned 6
contributors 9
BRIEFING FOR NEW SESSION
What changed since your last pull
02
pattern
Repository layer owns every database transaction
Source: pull request #214
03
bugfix
Refresh tokens rotate atomically under load
Source: auth/session.ts
Claude Code
Cursor
Windsurf
LOCAL FIT
Why Seoul engineering teams choose MemHub
In Seoul, software teams ship faster when their tools share one memory instead of three silos. New engineers ramp in days because precedent is searchable, not oral history.
Because writes land locally first and synchronize later, flaky connections never block the work.
- Seoul: connect Claude Code, Cursor, Windsurf, Cline, Codex CLI, or Antigravity to one shared memory. It also plugs into any CLI that speaks MCP or reads a JSON config.
- Local-first writes keep sessions productive through outages and travel.
- One command onboards the whole team: memhub login, then connect for Claude Code, Cursor, Windsurf, Cline, Codex CLI, or any JSON client.
UNIVERSAL COMPATIBILITY
Plugs into any CLI or development tool you use.
If it speaks MCP or reads a JSON config, MemHub plugs in. One command wires the majors; everything else joins as a generic client.
- Claude Code
- Cursor
- Windsurf
- Cline
- Codex CLI (GPT)
- GitHub Copilot
- Google Antigravity
- Gemini CLI
- Aider
- OpenCode
- CI bots via REST
WHAT YOU GET
Agent Memory Database in Seoul, without the trade-offs.
01
Vector-ready schema
pgvector HNSW indexes ship enabled — flip embeddings on when you want semantic search.
02
Row-level security
Tenant context policies mean a bug in one project’s code cannot read another org’s memories.
03
Operations as usual
Backups, replicas, point-in-time recovery — reuse the PostgreSQL playbook you already run.
04
Migrations included
Versioned SQL migrations with checksums; upgrade path is a command, not a migration essay.
05
Query it directly
BI tools, notebooks, and audits connect over standard SQL — no proprietary export formats.
06
PostgreSQL native
One battle-tested engine for memory, sync cursors, audit events, and billing metadata alike.
The honest comparison.
No feature-matrix gymnastics — just what changes the week you adopt MemHub.
| Status quo | With MemHub |
|---|---|
| Context lives in heads and DMs | Memory is structured, shared, and searchable |
| Every agent re-learns the repo | Sessions open pre-briefed with only what changed |
| Rules files cap out at ~2k tokens | Unbounded versioned history stays one query away |
Connect the tools already in your repository.
One command injects configuration for Claude Code, Cursor, Codex CLI, Google Antigravity, Windsurf, Cline, or any JSON MCP client.
repository terminal
$ npm install --global @simplexlat/memhub
$ memhub login
✓ project resolved · context ready · memory synchronized
Questions, answered straight.
Can I run it managed?
Yes — any Postgres 16 with pgvector works, including Easypanel, RDS, Cloud SQL, or Neon; self-hosted and hosted modes share identical schemas.
Why PostgreSQL instead of a vector-only store?
Memory needs relations (tenants, projects, authors, versions) as much as vectors. PostgreSQL provides both, plus the operational maturity no niche store matches.
What happens when our connection drops?
Nothing interrupts the work. Writes commit to the local SQLite outbox and synchronize automatically when the shared service is reachable again.
Can we try it without sending data anywhere?
Yes — the local engine works fully offline, and the self-hosted stack keeps even the shared layer on your own PostgreSQL.
How much does it cost?
A Community tier is free, and the Team plan adds multi-project contexts, seats, and higher synchronization limits. Self-hosting is unlimited.
MORE LOCATIONS
Agent Memory Database in San FranciscoAgent Memory Database in New York CityAgent Memory Database in AustinAgent Memory Database in SeattleAgent Memory Database in BostonAgent Memory Database in ChicagoAgent Memory Database in DenverAgent Memory Database in Los AngelesAgent Memory Database in AtlantaAgent Memory Database in MiamiAgent Memory Database in DallasAgent Memory Database in San Diego
EXPLORE OTHER USE CASES
AI Agent Memory
Give every AI coding agent a persistent, team-wide memory.
MCP Memory Server
A hosted MCP memory server your whole team connects to in one command.
Claude Code Memory
Claude Code memory that survives sessions and reaches the whole team.
Cursor Memory
Carry Cursor context across sessions, machines, and teammates.
Codex CLI Memory
Persistent context for Codex CLI — personal today, shared tomorrow.
Cross-Agent Memory
One memory behind Claude Code, Cursor, Codex CLI, and Antigravity.
Developer Knowledge Base
A living knowledge base that writes itself while your team codes.
Self-Hosted AI Memory
Run the entire memory stack on your own PostgreSQL — nothing leaves.
Coding Agent Context
Context engineering for coding agents — brief, don’t dump.
Engineering Decision Log
An ADR habit your team will actually keep — logged by the agents doing the work.
Windsurf Memory
Persistent Windsurf context shared with every other agent on the team.
Cline Memory
Give Cline a memory that survives tasks, machines, and teammates.
Copilot Team Memory
What Copilot’s custom instructions can’t hold, MemHub remembers.
Antigravity & Gemini Memory
Persistent context for Google Antigravity, Gemini CLI, and the agents around them.
THE NEXT SESSION STARTS HERE