memex
This is a persistent memory layer for AI agents that actually handles recency, not just semantic similarity. It exposes tools to store, search, and delete memories via MCP, backed by Postgres with pgvector and local ONNX embeddings. The search endpoint fetches 3x candidates by cosine similarity, then re-ranks them with exponential recency decay (configurable half-life) so your agent surfaces recent context alongside semantically strong matches. Runs entirely local via Docker Compose with no API keys. Load tests show 120 RPS sustained on modest hardware. Reach for this when your agent needs to remember user preferences, past conversations, or facts across sessions without collapsing everything into a single similarity score that ignores time.
- tools
- 4
- transport
- HTTP
- last updated
- Jun 4, 2026
Tools
4 toolsPublic tool metadata for what this MCP can expose to an agent.
store_memoryEmbed + persist a memory (type, importance configurable)
Embed + persist a memory (type, importance configurable)
No parameter schema in public metadata yet.
search_memoriesSemantic + recency ranked retrieval with configurable alpha
Semantic + recency ranked retrieval with configurable alpha
No parameter schema in public metadata yet.
delete_memoryForget a specific memory by UUID
Forget a specific memory by UUID
No parameter schema in public metadata yet.
count_memoriesHow many memories an agent/user pair has
How many memories an agent/user pair has
No parameter schema in public metadata yet.
Sponsors
Agent CaffeineSponsoredClose your lid, come back to finished work
SnugSponsoredThe clipboard manager for the AI eraPromote your product hereReach 125k+ visitors monthly
WaxSponsoredBlock anything with supervision, without erasing your iPhonePromote your product hereReach 125k+ visitors monthly
knobsSponsoredSee your site on any phone, any day, in any languageReadme
memex
A production-grade persistent memory service for AI agents. Agents forget everything between sessions by default — memex fixes that. It stores, retrieves, and ranks conversation memory using semantic search with recency decay, so agents surface what's relevant and recent, not just what's semantically closest.
Details
- Repository
- github.com/ayushagrawal288/memex
- Categories
- AI & LLM Tools, Cloud & Infrastructure, Search & Web Crawling
- Package
ghcr.io/ayushagrawal288/memex:v0.1.0- Transport
- HTTP
- Registry
- MCP registry, active
Related AI & LLM Tools MCP servers
View all
Conductor Graph Mcpbase76-research-lab/conductor-graph-mcpExposes a CognOS agent system as a machine-readable graph. Full system snapshot with one tool call.0stars
Mcp Grantiqbch1212/mcp-grantiqSearch US grants + federal contracts (Grants.gov + SAM.gov) from any LLM.0stars
Dear Userbleedmode/dearuserTells you how you and your Claude agent actually work together. Local-only, no API keys.0stars
Quarterbackbobbyrgoldsmith/quarterbackStrategic task prioritization and agent orchestration for multi-project operators0stars
Gonka Second Opinion (multi-model)bystray/gonka-mcp-serverA second opinion for AI agents: one prompt across several live Gonka models + roles, one call.0stars
Doomcapitalthought/doomObsolescence-risk screening: predicts if a product gets absorbed into core LLM-platform features.0stars