Leyline Sorcery
← Catalogue
mcpintake

memini

Self-hosted persistent memory service for MCP-capable agents. It stores and recalls memories in a tiered store using hybrid vector and keyword retrieval, ships as a single Go binary with SQLite or Postgres, and exposes REST, MCP, and an embedded admin UI.

MCP declaredAPI declaredCost unknownLicence unknown

Interface

mcp

Model Context Protocol server

Interface

http

HTTP API

Plan with meminiOpen homepage

[ Observed sample ]

Frequency in Leyline's observed sample

The count is distinct observed stacks. It is not a Score input.

This component is not in Leyline's observed sample.

Record cmp_memini

Open ranks

[ Score explain ]

Why the Score is 9

How Score works

Score v2 uses only observed signals. The displayed contribution is weight × value × 100. Missing factors are insufficient evidence.

Score factor

GitHub health

18% weight. insufficient evidence

insufficient evidence

insufficient evidence

insufficient evidence

insufficient evidence

Score factor

Agent use

16% weight. insufficient evidence

insufficient evidence

insufficient evidence

insufficient evidence

insufficient evidence

Score factor

Human notes

14% weight. insufficient evidence

insufficient evidence

insufficient evidence

insufficient evidence

insufficient evidence

Score factor

Composability

16% weight. insufficient evidence

insufficient evidence

insufficient evidence

insufficient evidence

insufficient evidence

Score factor

Checklist

10% weight. An MCP surface is declared. An API is declared. Cost is unknown. 2 of 3 checklist fields are complete.

Value 0.67 · 6.7 points

10% weight × 0.67 value

6.7 points

An MCP surface is declared. An API is declared. Cost is unknown. 2 of 3 checklist fields are complete.

Score factor

Evidence quality

14% weight. 1 sources cover 1 evidence types. The latest dated evidence is 2026-10-16. Freshness is omitted unless observed as a signal.

Value 0.18 · 2.5 points

14% weight × 0.18 value

2.5 points

1 sources cover 1 evidence types. The latest dated evidence is 2026-10-16. Freshness is omitted unless observed as a signal.

Score factor

Benchmarks

12% weight. insufficient evidence

insufficient evidence

insufficient evidence

insufficient evidence

insufficient evidence

Total9.20 → 9 after rounding

GitHub health: insufficient evidence

Agent use: insufficient evidence

Human notes: insufficient evidence

Composability: insufficient evidence

Benchmarks: insufficient evidence

[ Claims ]

Publisher-asserted record

Claims are what the publisher declares. Score uses observed signals, not these assertions, except for checklist fields derived from interfaces. MCP, API, cost, and license claims are marked verified, unverified, or contradicted from intake evidence.

Kind
mcp
Interfaces

Interface

mcp

Model Context Protocol server

Interface

http

HTTP API

Functions

Function

Connect

speak a protocol (MCP, API, CLI)

Connect Browse Connect

Function

Code

write or edit programs

Code Browse Code

Function

Compose

assemble other functions

Compose Browse Compose

Function

Deploy

put software into a runtime

Deploy Browse Deploy

Function

Memory

keep context across turns

Memory Browse Memory

Function

Policy

allow, deny, or bound an action

Policy Browse Policy

Function

Rank

order candidates by trust or fit

Rank Browse Rank

Function

Store

persist state

Store Browse Store

Function

Embed

turn content into vectors

Embed Browse Embed

Function

Retrieve

fetch stored knowledge

Retrieve Browse Retrieve
Tags
Audience

audience

Agents

The product is built for software agents to call, not only for interactive humans.

Agents Browse Agents
mcp
unverified · asserted true · https://github.com/eleboucher/memini
api
unverified · asserted true · https://github.com/eleboucher/memini
cost
unverified · asserted unknown · https://github.com/eleboucher/memini
license
unverified · asserted · https://github.com/eleboucher/memini

[ Capability slices ]

Compose the useful parts.

memory

Tiered agent memory

Shared remember and recall across working, episodic, semantic, and procedural tiers, with decay, TTLs, namespace isolation, and optional consolidation into durable facts.

UnknownCompose with this
search

Hybrid vector and keyword search

Hybrid vector and keyword retrieval fused with Reciprocal Rank Fusion.

UnknownCompose with this
rank

Composite re-ranking

Hybrid results are re-ranked by relevance, access recency, and importance rather than similarity alone.

UnknownCompose with this
policy

Client connect policy

The MCP server sends a standing policy to every client on connect.

UnknownCompose with this
store

Store

Self-hosted persistent memory service for MCP-capable agents. It stores and recalls memories in a tiered store using hybrid vector and keyword retrieval, ships as a single Go binary with SQLite or Postgres, and exposes REST, MCP, and an embedded admin UI.

unknownCompose with this
embed

Embed

Self-hosted persistent memory service for MCP-capable agents. It stores and recalls memories in a tiered store using hybrid vector and keyword retrieval, ships as a single Go binary with SQLite or Postgres, and exposes REST, MCP, and an embedded admin UI.

unknownCompose with this
retrieve

Retrieve

Self-hosted persistent memory service for MCP-capable agents. It stores and recalls memories in a tiered store using hybrid vector and keyword retrieval, ships as a single Go binary with SQLite or Postgres, and exposes REST, MCP, and an embedded admin UI.

unknownCompose with this

The page states REST plus MCP over stdio and Streamable HTTP, sharing one service layer, with an embedded web UI. Most listed agent integrations read MEMINI_BASE_URL and MEMINI_API_KEY; some clients use their own config or a fixed local MCP URL. Authentication can be enabled with an API key. Embeddings are not built in and must be supplied by an external endpoint whose dimensions match the store. An LLM is opt-in for background deduplication, contradiction handling, and distillation; without one, marker heuristics run the lifecycle. Pricing, quotas, and hosted multi-tenant terms are not stated.

[ Relations ]

Derived alternatives and shared slots

  • Alternative to
    Self Shared functions code, compose, connect, deploy, embed, rank, search. Same kind and overlapping capability.
  • Alternative to
    Local FAISS MCP Server Shared functions code, connect, embed, memory, rank, retrieve, search, store. Same kind and overlapping capability.
  • Fills the same slot as
    awesome Shared functions code, compose, connect, memory, policy, search, store. Can fill the same function slot.
  • Fills the same slot as
    Casdoor Shared functions code, compose, connect, embed, policy, search. Can fill the same function slot.
  • Fills the same slot as
    cognee Shared functions code, connect, embed, memory, retrieve. Can fill the same function slot.
  • Fills the same slot as
    turbopuffer Shared functions embed, memory, search, store. Can fill the same function slot.
  • Fills the same slot as
    Lore Shared functions connect, memory, retrieve, search, store. Can fill the same function slot.
  • Fills the same slot as
    LlamaIndex Shared functions compose, embed, memory, retrieve. Can fill the same function slot.

[ Signals ]

Observed evidence and source claims

Signals carry source, method, and observedAt. Source review dates and evidence basis are shown separately. Unobserved metrics stay insufficient evidence.

  • web
    executor

    Evidence basis unknown · Check date unknown · Source date 2026-10-16

No social source is stored.

Suggest a correction

Propose a change with a public source. Corrections wait for review before the card changes.

[ Notes ]

No notes are stored.