Skip to main content
Glama
deenesjakoruzh
FireCulex

int

by FireCulex
OverviewSchemaSourceRelated ServersDiscussionsAdmin
Python
Remote

int

Self-hosted, Dockerized, open-source AI memory system. Exposes a project-scoped memory store to AI coding assistants via the Model Context Protocol (MCP), so assistants can recall prior-session learnings without re-running expensive discovery work.

Status

v1 in progress. Spec at docs/spec.md; intent at docs/intent.md; task breakdown at tasks/todo.md.

Related MCP server: AIVectorMemory

Quick start

cp .env.example .env          # fill in API_KEY and GEMINI_API_KEY
docker compose up -d          # server + Qdrant
docker compose logs -f int    # tail server

Point OpenCode at the server by adding a remote MCP entry to your opencode.json:

{
  "$schema": "https://opencode.ai/config.json",
  "mcp": {
    "int": {
      "type": "remote",
      "url": "http://localhost:8000/mcp",
      "headers": { "API_KEY": "{env:API_KEY}" }
    }
  }
}

See docs/deployment.md for full setup, CLI usage, and common pitfalls.

Tools

Tool names are bare (add, search, ...). The MCP server is registered as int, so MCP clients expose them as int_add, int_search, int_list, int_delete.

Tool

Inputs

Output

add

project, type, content

memory_id

delete

memory_id

deleted: bool

search

project, query, limit=5

ranked SearchResult[]

list

project

metadata-only Memory[]

Stack

Python 3.14, FastAPI + MCP (mcp Python SDK, Streamable HTTP), Qdrant (separate container), gemini-embedding-001 (L2-normalized to 768 dims via MRL), Docker Compose, uv for dep management.

License

Apache-2.0 — see LICENSE for details.

This server cannot be deployed

One endpoint for every MCP toolRayrun runs every MCP server behind one URL and holds the credentials. Each call is decided per client, tool, and argument, then recorded.ray.runAd

Maintenance

ActivityMaintained
ResponsivenessNo issues

Resources

Related MCP Connectors

Related MCP Servers

  • Mono Memory MCP

    potato-castle
    A
    license
    Not graded
    quality
    D
    maintenance
    A self-hosted MCP server that provides AI assistants with a shared, persistent SQLite-backed memory for storing and retrieving project context, decisions, and discoveries. It enables cross-session continuity and team-wide knowledge sharing to keep AI coding tools aligned and informed.
    3
    MIT

  • AIVectorMemory

    Edlineas
    A
    license
    B
    quality
    D
    maintenance
    MCP server that provides cross-session persistent memory for AI coding assistants using local vector database and semantic search, enabling automatic recall of project context, issues, and tasks.
    9
    91
    Apache 2.0

  • Project Brain

    myczh-1
    A
    license
    Not graded
    quality
    D
    maintenance
    A local MCP memory server that gives AI assistants durable project memory across coding sessions, storing context, changes, and decisions.
    0
    1
    MIT

  • Local Context MCP

    zwinglio
    F
    license
    Not graded
    quality
    B
    maintenance
    A self-hosted AI memory service that shares durable project context between MCP-compatible clients such as Claude Code and Devin, via a REST API and a Streamable HTTP MCP server with tools to remember, recall, list, and forget memories.
    -