Memdoor

Memdoor is the coding agent for durable workflows automates repeatable tasks from CI, schedules, and shells.

Memdoor is an open source coding agent built for durable, repeatable workflows. It operates headless or interactive, launched from CI, schedules, or shells, and connects to various AI providers using your own API key or subscription. A built-in decision model reduces token usage and filters irrelevant data before sending prompts to the chat model. Users define multi-step workflows as YAML files that execute as graphs in the repository, supporting parallel steps and manual approval gates. It features local file editing, test execution, build fixing, and model management through terminal commands or editor integrations. Data stays local on the machine except for model requests sent directly to connected providers.

Memdoor at a glance

Best for: Software engineer
Runs on: macOS, Windows, Linux, CLI, VS Code, JetBrains, Zed, Web, self-hosted, open source
Pricing: Free plan; Pro $10/month; Enterprise available
Released: October 2026
Category: AI coding tool
Website: github.com/guregodevo/memdoor

Key features

Decision model filtering for token reduction
YAML task files for multi-step workflows
Local gateway running on machine
Approval gates for workflow steps
File reading, patch writing, and build fixing
Provider connection for custom keys
Interactive terminal window and commands

Use it for

Running automated coding workflows from CI or cron
Executing multi-step tasks with approval gates
Fixing code and running builds headlessly
Managing coding agents from an editor or terminal

Alternatives to Memdoor

OpenCode — Works with 75+ model providers, local models, and existing subscriptions
claude-mem — It shares persistent, searchable agent memory across sessions and devices.
ProjectMem — Records project history locally to prevent AI from repeating failed work.
Emdash — Run multiple coding agents in parallel with full worktree isolation.
Memdebug — Tracks AI agent memory in a tamper-evident ledger and rolls it back safely.
Hazzel — Brings inspectable safety mechanisms, local model support, and zero subscriptions.
OpenClaude — Connects one terminal interface to over 200 different AI models.
Weft — It enables safe parallel coding agent work without file locks.
Archon — Orchestrate multiple AI agents with isolated workflows via a single command layer.
Alcubi — Run multiple coding agents concurrently without managing live sessions.
Command Code — Repairs open model tool calls and learns developer patterns automatically.
memsearch — It shares continuous context across Claude Code, Codex, and other AI coding agents.

All Memdoor alternatives