Kody launches as a Fair Source home giving AI agents shared memory, secrets and packages
DavidWells · x · 2026-09-20
David Wells shared Kody, a platform that gives your existing AI agents (Claude, Cursor, Codex, etc. via MCP) durable shared state.
Key features:
- Memory: account-level persistent facts surfaced to every connected agent
- Secrets: API keys agents can use but never read (no secretget by design)
- Packages: reusable code saved by agents, callable from any agent
- Jobs, workflows, webhooks that run packages on schedules or triggers without a model in the loop
- Integrations/MCP servers plus an inbox — each account gets an email address that can kick off work
Kody is Fair Source (readable, forkable, self-hostable; converts to Apache 2.0 after two years), runs no chat-model loop and bills no chat tokens.
Related event: Kody Launches Shared Persistent Memory and Skill Hub for AI Agents(2 posts)→
More from coding & agent
- Multi-agent systems work best with clear roles, not more agents — _jaydeepkarale · 2026-09-20
- Jev founder: all AI models are built for human-in-the-loop, not true automation — hardimanjames · 2026-09-20
- Karpathy's Stanford AI engineering lecture: from LLM to Graph in five steps — irinarish · 2026-09-20
- Browser Use's open-source Jev agent searches Google Flights in 7 seconds for $0.0039 — FinanceYF5 · 2026-09-20
- If Jev makes your evals so much cheaper, why was your judge so expensive? — zeeg · 2026-09-20
- Benchmark Reality Gap: Why High-Scoring Agents Still Fail in Production — AIpro96 · 2026-09-20