caveboy/mnemoforge repository overview

⁠SloplessCode

Operational continuity infrastructure for AI coding agents.

SloplessCode helps AI coding agents continue real project work across IDE restarts, model switches, context loss, interrupted sessions, and multi-agent handoffs. It stores project state, task lifecycle, checkpoints, governed rules, decisions, and searchable memory behind HTTP and MCP interfaces.

Formerly Mnemoforge. During the rename transition, the same image is also published as caveboy/mnemoforge:latest for compatibility.

⁠What It Is

SloplessCode is not just chat memory. It is an operational runtime for agent-led software development:

• project memory and semantic search
• task and improvement lifecycle
• checkpoints and handoff notes
• project laws/rules and operational guidance
• MCP tools for coding agents and local models
• backup/restore and storage-health workflows

The goal is simple: an agent should not start from zero every time a session is reset. It should be able to ask the system what happened, what matters, and what the next safe action is.

⁠Relevant Categories

• Machine learning & AI
• Developer tools
• Databases & storage
• API management
• MCP tooling for AI coding agents
• Local-first project memory and operational continuity

⁠Quick Start

The recommended user path is Docker Compose from the repository:

git clone https://github.com/Utundry/sloplesscode.git
cd sloplesscode
python scripts/configure_public.py --non-interactive
docker compose --env-file .env.user -f docker-compose.user.yml up -d

Health check:

curl http://localhost:8000/api/v1/health

MCP SSE endpoint:

http://localhost:8000/mcp/sse

⁠Image Tags

Current image:

docker pull caveboy/sloplesscode:latest

Compatibility image during the rename transition:

docker pull caveboy/mnemoforge:latest

Immutable commit tags are published alongside latest when release publishing uses --tag-current-git-sha.

⁠Data Persistence

Container updates should not delete your project data if you keep the Compose volumes/directories intact.

Do not run destructive volume cleanup commands unless you have a backup:

• avoid docker compose down -v
• avoid docker system prune --volumes
• do not delete the local data directory or named volume

SloplessCode separates source-of-truth system data from Qdrant semantic index data:

• sloplesscode_system_data stores SQLite/JSON project memory, tasks, laws, aliases, checkpoints, and lifecycle state.
• mnemoforge_qdrant_data stores Qdrant semantic index files for compatibility with the rename transition.

Qdrant is treated as rebuildable semantic indexing data; system data is the critical backup target.

⁠Alpha Status

SloplessCode is alpha software under active architectural evolution. MCP behaviors, workflows, and storage surfaces may change. Public images are tested through Docker-backed checks and live MCP smoke checks before publishing, but you should still keep backups and report regressions.

⁠Useful Links

• GitHub: https://github.com/Utundry/sloplesscode
• Docker Hub: caveboy/sloplesscode
• Compatibility Docker Hub image: caveboy/mnemoforge
• Maintainer: Nikolay Laptev, [email protected]