LuizEduPP/rememb

AI agents forget everything between sessions. rememb gives them persistent memory — local, portable, and works with any agent.

---

The problem

Every dev using AI professionally hits this wall:

Session 1: "We're using PostgreSQL, auth at src/auth/, prefer async patterns."
Session 2: Agent starts from zero. You explain everything again.
Session 3: Same thing.

Existing solutions (Mem0, Zep, Letta) require servers, API keys, and cloud accounts.
You just want the agent to remember your project.

---

Install

pip install rememb

---

Quick Start

With MCP (recommended)

Zero friction. No CLI commands. Native IDE integration.

1. Add to your IDE's MCP config:

{
  "mcpServers": {
    "rememb": {
      "command": "rememb",
      "args": ["mcp"]
    }
  }
}

2. Restart your IDE.

The agent now automatically reads memory at session start, writes when learning something new, and searches when needed.

If you want multiple MCP clients on the same machine to reuse one already-running rememb process, start a persistent local SSE transport:

rememb mcp --transport sse --host 127.0.0.1 --port 8765

This keeps one MCP process alive, so repeated clients can hit the same loaded embedding model through http://127.0.0.1:8765/sse and http://127.0.0.1:8765/messages/.

Do not put --transport sse inside a stdio MCP client config. stdio clients expect JSON-RPC on stdin/stdout; the SSE mode exposes an HTTP endpoint and must be started separately.

Local usage without MCP

rememb          # Open the TUI
rememb fetch-model   # Download the local embedding model for semantic search

---

How it works

.rememb/
  entries.json   ← structured memory (project, actions, systems, user, context)
  meta.json      ← project metadata
  config.json    ← limits, sections, TUI behavior, semantic model settings

A JSON file in your project. Your agent reads it at the start of every session.

User: "We're using PostgreSQL, auth at src/auth/, async patterns"
Agent: [rememb_write] → Saved

[New session]
Agent: [rememb_read]  → Context loaded
Agent: "I see you're using PostgreSQL with auth at src/auth/..."

Search uses local semantic embeddings (no API, no cloud). The embedding model is unloaded after a short idle window by default, so the process does not keep the full model resident forever.

rememb now writes the full configuration set to .rememb/config.json during initialization, so all supported knobs live in one place:

{
  "max_content_length": 1000000,
  "max_tag_length": 500,
  "max_tags_per_entry": 100,
  "max_entries": 100000,
  "sections": ["project", "actions", "systems", "requests", "user", "context"],
  "section_icons": {
    "project": "◈",
    "actions": "↯"
  },
  "section_colors": {
    "project": "#d84848",
    "actions": "#d08020"
  },
  "entry_batch_size": 24,
  "entry_load_threshold": 6,
  "semantic_model_idle_ttl_seconds": 15,
  "semantic_model_name": "paraphrase-MiniLM-L3-v2"
}

Set semantic_model_idle_ttl_seconds to 0 to unload the model immediately after each semantic operation. If you want a smaller model, you can switch semantic_model_name to another SentenceTransformers model such as paraphrase-MiniLM-L3-v2.

entry_batch_size and entry_load_threshold control how aggressively the TUI lazy-loads cards from the local store.

Section names are normalized to lowercase, duplicates are ignored after normalization, and removing a section with existing entries automatically migrates those entries to uncategorized. meta.json is kept in sync with the current effective section list.

Environment overrides are also available: REMEMB_SEMANTIC_MODEL_IDLE_TTL_SECONDS and REMEMB_SEMANTIC_MODEL_NAME.

---

Memory sections

Section	What to store
project	Tech stack, architecture, goals
actions	What was done, decisions made
systems	Services, modules, integrations
requests	User preferences, recurring asks
user	Name, style, expertise, preferences
context	Anything else relevant

---

TUI

rememb includes a full terminal UI built with Textual.

rememb          # Open the TUI

Features:

• Grid of memory cards — browse all entries organized by section
• Sidebar navigation — filter by section with entry counts
• Inline search — press / to search across all entries
• Tag filter — click a tag pill to combine exact tag filtering with the current text search
• Side panel — create or edit entries without leaving the screen
• Full config screen — edit sections, section icons, semantic model, limits, and lazy-loading behavior with F2
• Dynamic layout — grid adapts to terminal width (1–4 columns)
• Keyboard shortcuts — Ctrl+N new, Ctrl+R refresh, / search, Q quit

Cards keep their content preview and timestamps, but tag rendering is intentionally capped so entries with many tags do not grow unbounded in the grid.

New custom sections get a random color automatically and start with a generic icon until you customize them in the config screen.

The semantic search MCP tool also accepts an optional exact tag filter, so IDE clients can restrict semantic matches before ranking.

---

CLI

rememb          # Open the TUI
rememb mcp      # Start MCP server for AI agent integration over stdio
rememb mcp --transport sse --host 127.0.0.1 --port 8765   # Start one persistent local MCP process
rememb --version, -v    # Show version
rememb --help, -h       # Show help

---

Compatibility

The current compatibility surface is tracked explicitly in COMPATIBILITY.md.

Short version:

• Python 3.10 to 3.12 are covered by CI
• CLI contract and MCP tool schema have automated test coverage
• stdio MCP is the primary documented integration path
• SSE MCP is documented, but not yet covered by end-to-end automated client tests
• release automation and Trusted Publishing are documented in RELEASE.md

---

Design

• Local first — plain JSON file in your project
• Portable — copy .rememb/ anywhere, it works
• Agnostic — any agent, any IDE (MCP or CLI)
• No lock-in — no servers, no API keys, no accounts

---

Contributing

git clone https://github.com/LuizEduPP/Rememb
cd rememb
pip install -e ".[dev]"

PRs welcome. Issues welcome. Stars welcome. 🌟

---

License

MIT