tashfeenahmed/scallopbot

What is it?

Key features

• 01Memory that decays, fuses and archives over time, with recall mixing BM25 and embeddings.
• 02Cost-aware routing across seven LLM providers with failover and daily/monthly budgets.
• 03Two-way MCP support: consumes MCP servers and exposes memory tools to other clients.
• 04Background nightly consolidation, self-reflection and proactive gap scanning.
• 05Compatible with OpenClaw and agentskills.io style skills, with 34 bundled skills.
• 06Local voice transcription and synthesis options, plus a web dashboard installable as an app.

Use cases

• Running a private always-on assistant on a home server or Raspberry Pi that you reach through Telegram.
• Sharing one persistent memory between your chat assistant and coding tools like Claude Code via MCP.
• Keeping LLM spend predictable by routing requests across cheaper and stronger models with budget caps.
• Experimenting with agent memory designs that go beyond simple note storage.

Quick start

1. 1

   Clone the repository

   git clone https://github.com/tashfeenahmed/scallopbot.git
2. 2

   Enter the folder and install dependencies

   cd scallopbot && npm install
3. 3

   Copy .env.example to .env and add at least one LLM provider API key, then build

   npm run build
4. 4

   Start the bot; the dashboard runs at http://localhost:3000

   node dist/cli.js start

Things to know

• !Requires Node.js 24 or newer.
• !Needs at least one paid LLM provider API key to run.
• !Not published to the npm registry yet, so it must be built from a clone.

Topics

FAQ

What is scallopbot?

Self-hosted personal AI assistant with bio-inspired memory that is fused, linked and forgotten over time, exposed over MCP and compatible with OpenClaw skills. ScallopBot is a personal AI agent that runs on your own server and talks to you through Telegram, a web dashboard, a CLI, or channels such as Discord, Slack, WhatsApp, Signal and Matrix. Its memory lives in a single SQLite file and combines BM25 keyword search with embeddings and optional LLM re-ranking. Memories decay by category, get fused into summaries, and low-utility items are soft-archived before pruning.

Is scallopbot free to use? What license does it have?

scallopbot is open source under the MIT license. Check the license terms before using it in commercial products.

Is scallopbot still maintained?

Actively maintained. The last commit to scallopbot was on Oct 16, 2026.

What can I use scallopbot for?

Running a private always-on assistant on a home server or Raspberry Pi that you reach through Telegram. Sharing one persistent memory between your chat assistant and coding tools like Claude Code via MCP. Keeping LLM spend predictable by routing requests across cheaper and stronger models with budget caps. Experimenting with agent memory designs that go beyond simple note storage.

How do I install scallopbot?

Clone the repository: git clone https://github.com/tashfeenahmed/scallopbot.git Enter the folder and install dependencies: cd scallopbot && npm install Copy .env.example to .env and add at least one LLM provider API key, then build: npm run build Start the bot; the dashboard runs at http://localhost:3000: node dist/cli.js start

Where can I get scallopbot?

The source code and installation instructions are on GitHub: https://github.com/tashfeenahmed/scallopbot