myna

Myna

The system for everything you're trying to keep in your head. Nothing leaves your machine.

AI Chief of Staff for tech professionals — your projects, your team, your meetings, your email, all running locally inside Claude Code.

Built autonomously by Claude Code, end to end — 12 agents, 31 skills, no vibe-coding. How it was built →

The day you actually have

It's 7:45 AM. You haven't opened your laptop, and you already know roughly what's waiting. A 9 AM 1:1 with someone who's been stretched thin — and you still haven't written down the pattern. Six meetings stacked back to back, none of them prepped. An inbox of 38 unread, four of which are real and the rest of which look real. A VP who wants a "quick Phoenix risk note" by Friday, which means remembering what happened in week one of Phoenix — 47 days ago. Two people owe you something; you owe four people something back. And somewhere in there, you're supposed to do your own work.

You're not behind. This is just the job. If you manage projects and people, most of your day isn't decision-making — it's information management: triage, prep, catch-up, drafting the same kind of update for the fifth time this month, trying to recall what was decided three weeks ago. Roughly 80% of the week is moving information around. The other 20% is the work only you can do — and it's the part you actually get paid for.

The tools you already have don't fit the shape of that 80%:

• General AI chat (Claude, ChatGPT, Gemini) starts every session from zero. It doesn't know who Marcus is, what Phoenix is blocked on, or how you sign your emails — so you spend the first three minutes re-explaining context, and none of it survives the closed tab.
• Local AI notebooks (Khoj, Fabric, PrivateGPT) answer questions about notes you already wrote. They don't do the work — no prep brief, no triage routing, no drafted escalation. They're search engines with a chat box, not assistants.
• Cloud "AI assistants" for work want to live in the middle of your stack and ingest your email and calendar onto someone else's servers. At most companies that's a non-starter — and even where it's allowed, it's one more vendor to evaluate and one more thing to break.

Myna sits exactly in that gap.

What Myna is

Myna is a Claude Code agent with 31 skills for the work that fills your week: it preps your meetings, triages your inbox, keeps every project and person up to date, and drafts your replies and status updates — all from prompts you type in plain language. It connects to the tools you already use through MCP, reading from your email, Slack, and calendar, and maintains a local knowledge base of your work as plain markdown: projects, people, meetings, decisions, action items, and history.

Plenty of tools can draft an email or summarize a thread. What sets Myna apart is the context behind every one: it works from that knowledge base instead of a blank slate. Install it once and every session builds on the last — the brief on Sarah this morning already includes the recognition you logged last week and the design review she led on April 9, and the Phoenix risk note you draft today knows the blocker is 20 days old, because Myna wrote it all down when it happened. Its drafts sound like you and reflect what's actually going on.

It drafts but never sends. It organizes but never decides. It surfaces but never hides. You still pick up the phone, send the email, and make the call — Myna just makes sure that when you do, you're not winging it from memory.

See it in action

A day in the life of an engineering manager using Myna. Every line is a prompt you'd actually type into Claude Code.

7:45 AM — coffee, laptop open
> daily-brief
↳ Daily note created. Today's meetings as a table with linked prep.
  Phoenix blocker flagged. Sarah Mitchell's reply overdue.

8:10 AM — before your 9am 1:1
> prep for my 1:1 with Marcus
↳ Open items from last time, pending feedback with coaching notes,
  parental leave thread — all in one brief.

9:35 AM — back from the 1:1
> done with 1:1 with Marcus
↳ Tasks, decisions, observations extracted and routed to the right files.

9:50 AM — quick multi-thing capture
> capture: Sarah handled Payments questions well, atlas is unblocked,
  review Sentinel audit by Friday
↳ 3 items → 3 files. Recognition, timeline update, task with due date.

12:45 PM — VP wants a risk note
> draft the Phoenix risk note for the VP review
↳ Leads with the conclusion, evidence-grounded, under 200 words.
  Ready for you to review, then send yourself.

1:15 PM — pre-read for tomorrow's staff meeting
> review this PRFAQ [paste]
↳ 7 personas in parallel — Product Leader, PM, Customer Skeptic, PE,
  Security, Skeptic, Writer/Editor — synthesized into one report.
  Saved to Reviews/2026-05-26-pricing-prfaq.md.