Meetless
Meetless: Active source of truth for your coding agents - watches sessions, captures decisions, and steers agents.
Last verified:
What is Meetless?
Meetless: MLA is a governance layer for AI coding agents that captures decisions during development sessions, requires human approval before they become project truth, and steers agents with current approved context before they act. Built for solo developers and teams using Claude Code and Codex, it solves 'context rot' by ensuring agents act on current decisions rather than stale or conflicting information.
Meetless pricing
Pricing model: Free
Free during beta
Meetless pros
- Human-in-the-loop approval workflow prevents unapproved agent decisions from entering the codebase
- Conflict reconciliation automatically detects and routes contradicting decisions to humans rather than letting agents guess
- Prepends current approved decisions to agent prompts, ensuring decisions persist across sessions without staleness
- Works across Claude Code and Codex with published benchmarks against CLAUDE.md and RAG approaches
Meetless cons
- Limited to Claude Code and Codex agents; no support for other AI coding tools yet
- Still in beta phase with no permanent pricing announced
- Requires active human review to approve captured decisions, adding overhead to the development workflow
Frequently asked questions about Meetless
Does MLA stop agents from making decisions?
No. Humans remain the architect—nothing an agent proposes becomes project truth until a human approves it.
Why is bigger context and better memory not enough?
The problem is not memory; it's authority. Larger context windows and memory systems can include both old and new decisions without clarifying which one is currently in force.
How does MLA prevent agents from reviving rejected approaches?
MLA captures decisions during sessions, requires human approval, reconciles conflicts with history, and prepends approved decisions to the agent prompt before it acts.
What happens when decisions conflict?
MLA preserves the history and shows both claims with the sessions they came from. Unresolved conflicts are routed to a person for decision, not left for the agent to guess.