Local chat sharing live
Start a conversation on your laptop, finish it on your desktop — or hand it to a teammate mid-thought.
Tokenly is a one-person company building developer tools in the open. Its flagship is Mako — a CLI that sits between your machine and your coding agent, so your team's Claude Code chats travel with you today, and policies and token limits travel with every request tomorrow.
Mako proxies the traffic between you and your agent's backend — and that's where your team's rules will live. Today it shares your local Claude Code chats.
/local-chat-sync uploads this machine's Claude Code transcripts, byte for byte.
Each transcript is sealed with AES-256-GCM before it is written, scoped to your team.
/local-chat-list shows what you and your teammates have synced, from any machine.
/local-chat-import writes it back where Claude Code expects — any unambiguous id prefix works.
claude --resume picks the conversation up exactly where it left off.
Chat sharing is the first thing Mako does. It won't be the last.
Start a conversation on your laptop, finish it on your desktop — or hand it to a teammate mid-thought.
Allow or block models and cap max_tokens before a request ever reaches the agent's backend.
Per-team and per-person budgets that stop a runaway agent loop before it becomes a bill.
See who used what, on which model, with p50/p95/p99 latency — without logging a single prompt.
Each one is small, does one job well, and is built in public.
The fastest way your team works with Claude Code — chat sharing now, policies and token limits next.
Read more →Voice activity detection, streaming speech recognition and a local LLM — wired end to end on your own hardware.
Visit Kural →Re-angles the resume you already have to the job you're looking at — cited, verified, one page.
Visit fast-CV →Tokenly is a one-person company. Every product starts as a public repository — follow along, open an issue, or just say hi.
Follow on GitHub