Founding 10 Challenge

10 spots 14-day honest outcome

Don't endorse us. Break us - or prove us right.

The first 10 people who run MemTether on their real AI-coding setup and submit an honest 14-day report get Founding Contributor status: your handle in the README, priority on feature requests, and a direct line to the maintainer.

What you do

Run it unchanged - pip install memtether && memtether init. No custom config.
Use it for 14 days - real work, real agents, real sessions.
Find what cannot be trusted - crashes, wrong memories, adapter failures, anything.
Submit any honest outcome - positive, negative, or mixed. All three count.

What we publish

Every submission (positive or negative) gets published in the README with attribution. If you find a real bug, you get credited as the finder. If everything works, you get credited as a verifier.

We do not cherry-pick positive reviews. Honest failures are more valuable than praise - they tell us what to fix.

Tracks

Track	What you test	Status
1. Claude Code	MCP config + cross-session recall	Open
2. Cursor	MCP config + multi-project memory	Open
3. Codex CLI	MCP + CLI remember/search	Open
4. Windsurf	MCP config	Open
5. Gemini CLI	MCP config	Open
6. Multi-client	Cross-client memory sharing	Open
7. Bi-temporal audit	Timeline / supersession chain	Open
8. Tenant isolation	Multi-tenant 0-leak	Open
9. Docker deploy	docker-compose + api_server	Open
10. Fresh machine E2E	Clean venv full workflow	Open

How to register

Open an issue at MemTether issues with title [Founding 10] <track> and your OS / Python / AI clients / start date. We confirm within 24h.

Rules

14 days from your start date. No extensions.
Submit via GitHub issue (public) - honesty is the point.
All 10 spots must complete within 60 days of this page going live, or the challenge resets.
Founding Contributors get README credit + issue priority. No payment, no equity.
Maintainer reserves the right to reject non-genuine attempts.