Local-first AI work companion

Kaisho knows your tasks, customers, time, notes and code — and runs your AI agents over them on schedule, with the model and API key of your choice. Use it from your desktop and plug it straight into Claude Code, Cursor, and Claude Desktop via MCP. Mobile sync and the hosted MCP gateway arrive with Companion in Q3. Your data stays on your machine.

Download for Desktop View Source

Up and running in three steps

Install the desktop app, point your AI at it, and let it remember.

1

Install

Grab the desktop app for macOS (Apple Silicon), Windows, or Linux from GitHub Releases. Or install via pip install kaisho && kai serve. Tasks, time, customers, notes and the KB all live in plain text on your machine.

2

Connect your AI

Add the Kaisho MCP server to Claude Code, Cursor, or Claude Desktop with a one-line config. Your AI can now read and write your tasks, look up a customer, log time, query the KB, and run shell commands you whitelist — without leaving the editor.

3

Let it run

Schedule cron-AI agents to do the boring parts: a Monday-morning briefing on last week, a Friday invoice draft, a daily KB digest. They run locally on your hardware (Hobby) or on our hosted token quota (Companion / Pro).

A memory layer for your AI — with a real productivity tool attached.

Kaisho is what your AI companion needs to be useful day-to-day: durable memory, scheduled runs, MCP-native integration with the editor you already use, and a real task / time / customer tool underneath.

A memory your AI can actually use

Tasks, time entries, customers, notes and a knowledge base — all in plain text, all on your machine, all queryable by your AI through MCP. Sessions in Claude Code or Cursor stop being amnesiac.

Scheduled AI agents

Cron-AI jobs run on your schedule, using the model and API key of your choice. A Monday-morning briefing on last week. A Friday invoice draft. A nightly KB digest. The agent reads your Kaisho data, does the work, leaves an artefact.

MCP-native

First-class MCP server. One-line setup for Claude Code, Cursor and Claude Desktop. Premium MCP integrations on Pro: Linear, GitHub Projects, Google Calendar, Slack — so your AI can read tasks from Linear, file a PR via the GitHub MCP, and book the time in Kaisho in a single chain.

Local-first by default

Hobby runs entirely on your machine — no account, no cloud, BYO API key. Companion and Pro add a hosted token quota and cross-device sync, but the source of truth stays your plain-text directory.

Bring your own model

Claude, GPT, Gemini, local Ollama, LM Studio, any OpenAI-compatible endpoint. Hobby is BYO key. Companion and Pro include a frontier-model quota so you can stop juggling keys — or keep your own for sensitive work.

Still a real productivity tool

Under the AI layer: drag-and-drop kanban, start/stop timers, per-customer contracts and budgets, unbilled-entry preview and CSV export for your accountant, weekly/monthly dashboards. The productivity tool is what gives the AI something to remember.

Every feature is also a kai command.

The web UI is optional. Every feature in Kaisho is a kai subcommand — and every kai subcommand is also an MCP tool. Start timers, query your KB, book time, run an agent — from the shell, from the editor, or from your AI session. Same surface, three entry points.

$
$
$
$
$
$

Never leave your editor.

kaisho-mode is an optional Emacs package that keeps your clock visible in the mode line and lets you start, stop, and resume timers without switching windows. It connects to the running kaisho backend via WebSocket — so clocks started in the web UI appear in Emacs instantly, and vice versa.

Works with any backend. No org-mode required. All operations go through the kai CLI and REST API.

~/.doom.d/packages.el
(package! kaisho-mode
  :recipe (:host github
          :repo "ridingbytes/kaisho-mode" ))
Mode-line & keybindings
Emacs ⏱ Beta GmbH 1:23  Kaisho
SPC n k k  — toggle clock
SPC n k s  — today's summary
SPC n k i  — insert backdated entry
Resume last clock on next toggle.
Reconnects automatically if the backend restarts.

Your data, human-readable.

Kaisho stores everything as plain text files you can open, edit, and version-control with any tool. Choose between org-mode and Markdown — the web UI and CLI work identically with both.

* [2026-04-10 Thu] [Acme Biotech]: Sample tracking
  :PROPERTIES:
  :SYNC_ID: 644e97a2-77a2-4ff8-b7d7-1c766ef3f2b8
  :UPDATED_AT: 2026-04-10T12:30:00.123456
  :TASK_ID: a1b2c3d4e5f6
  :CONTRACT: LIMS Phase 2
  :INVOICED: false
  :END:
  :LOGBOOK:
  CLOCK: [2026-04-10 Thu 09:00]--[2026-04-10 Thu 12:30] =>  3:30
  :END:

Your AI memory in your pocket

The mobile PWA gives you the same data your editor talks to: tasks, time, customers, dashboards. Chat to the advisor on the go — same model, same memory, same MCP context as on your laptop. Included on Companion and Pro.

Open Mobile App

Where your AI already lives.

First-class MCP for Claude Code, Cursor and Claude Desktop. Any OpenAI-compatible model on the back end. Plain-text storage you can pipe to anything else. No lock-in — on either side.

GitHub Ollama Claude API OpenRouter Emacs Org-mode Markdown LM Studio

Pick the level of memory you want for your AI.

Hobby is the open-source desktop. Companion, Pro and Team add a hosted memory layer, scheduled AI runs, and MCP integrations with the editor of your choice.

Start free. Pay only when you want your AI to run on a schedule and remember across sessions.

Hobby
Free
Open source. No account needed.
  • Full desktop app (macOS / Win / Linux)
  • CLI & local data, your machine only
  • Tasks, notes, time, customers, KB
  • Bring your own API key (any model)
  • Local AI via Ollama / LM Studio
  • Self-host the MCP server
View on GitHub
Companion
€29/mo
or €290/yr. Cancel anytime.
  • Everything in Hobby
  • 500k tokens / mo of frontier model
  • Cross-device sync + mobile PWA
  • Scheduled AI runs (cron agents)
  • MCP gateway for Claude Code / Cursor
  • Overage packs (€15 per 500k tokens)
Coming Q3 2026
Pro
€59/mo
or €590/yr. Cancel anytime.
  • Everything in Companion
  • 2M tokens / mo of frontier model
  • Premium MCP integrations
  • Linear, GitHub Projects, Calendar, Slack
  • Priority cron-AI compute
  • Priority email support
Coming Q3 2026
Team
€99/seat/mo
per seat / month. Minimum 2 seats.
  • Everything in Pro, per seat
  • Shared knowledge base
  • Team-wide cron agents
  • Role-based access control
  • Audit log
  • Annual invoice on request
Coming Q3 2026

No credit card for Hobby. Companion and Pro are personal-card priced — no purchase approval needed.

Common questions

Kaisho ships a Model Context Protocol server out of the box. It exposes your tasks, time entries, customers, notes and KB as tools your AI can call — add_task, book_time, list_customers, search_knowledge and so on. Add one line to your Claude Code or Cursor config and your AI session can read and write your Kaisho data without copy-paste.
Yes — all three are first-class targets with a one-line setup. See the MCP integration guide for the exact config per editor. The same MCP server works with any other MCP-compatible client (Zed and friends) over stdio. Companion (Q3) adds a hosted gateway so the same tools stay reachable when your laptop is closed.
Always. Hobby is bring-your-own-key by design — Claude API, OpenRouter, GPT, Gemini, local Ollama, LM Studio, any OpenAI-compatible endpoint. Companion and Pro add a hosted token quota of a frontier model so you can skip the key-juggling, but you can still point at your own provider for sensitive work.
Always on your machine. Kaisho is local-first — your tasks, time entries, customers, notes and KB live as plain text files in a directory you own. Companion and Pro add a hosted sync layer so the mobile PWA and your other devices stay in step, but the source of truth never leaves your laptop.
Scheduled jobs that run a prompt against your Kaisho data on a cron schedule. Today they run on your machine via APScheduler — bring your own model or API key. Companion (Q3) adds a hosted runner so jobs fire even when your laptop is closed, using the included token quota. Templates are included — weekly business pulse, Friday invoice draft, nightly KB digest. Write your own in YAML.
Yes — the kanban board, timers, customer/contract budgets, unbilled-entry preview and CSV export are all still there and still maintained. The repositioning is what we lead with, not what we removed. If you only want a time tracker, Hobby is enough and always will be.
Because the source of truth and the AI memory both sit on your machine. The hosted sync layer for Companion and Pro is a convenience for mobile and cross-device — not a database lock-in. You can cancel and keep your data, untouched, in the same directory you started with.

Open source by default. Hosted when you want it.

Kaisho is MIT-licensed. Run the desktop app, the MCP server, and the sync layer entirely on your own hardware — or pay for the hosted token quota and skip the setup. Either way, the source is on GitHub, the data is yours, and the architecture works without us.

View on GitHub
# Clone and run from source git clone https://github.com/ridingbytes/kaisho cd kaisho && pip install -e . kai serve
Kaisho mark

The Mark

In Japanese martial arts, kaisho (開手) means “open hand” — the position of readiness. The mark is an open bracket with three lines: the bracket is the palm, the lines are what you carry — tasks, hours, ideas.

The bracket is not closed. It accepts new entries, holds what matters, and expands when you need it to. Structure without rigidity.