[opensourcedex]

kition is a free, open source note taking & knowledge management project written in TypeScript and released under Apache-2.0. It has 520 GitHub stars, 39 forks and 2 open issues, and was last pushed 3 hours ago. On this registry it ranks #58 of 62 tracked projects in Note Taking & Knowledge Management, with 5 head-to-head comparisons available.

What is kition?

Kition is an Apache-2.0, local-first desktop workspace written in TypeScript and shipped as an Electron application, and it is aimed at people who want a tool-using AI agent to work directly inside their own Markdown documents, typed table files, and whiteboards instead of in a separate chat window.

What it is

Kition is a single desktop application that gathers several kinds of working surface into one place: Markdown documents with live preview, links, backlinks, callouts, code, math, diagrams, daily notes, search and export; structured table files that hold typed records, formulas, filters, groups, views, attachments and AI fields; and an infinite Whiteboard for diagrams, mind maps, flows, and freeform sketches. On top of these sits a tool-using AI agent that can search the web, inspect documents and table schemas, update content, and save its output into the active project, together with a visual workflow builder for trigger-and-action automations. The project lives in the note-taking and knowledge-management ecosystem, describes itself as local-first, and carries the notion-alternative topic tag in its repository metadata.

The concrete problem it addresses is the blank prompt. A capable agent still leaves the user to design the task, describe every input, choose an output format, and recover when a step goes wrong, and the resulting output usually has to be handed between disconnected tools by copy and paste. Kition instead gives the agent editable project files, typed records, attachments, and visible processes to work with, so a request starts from existing work rather than from nothing. When the agent edits the active Markdown document, Kition opens a document review surface that highlights additions, deletions, and rewrites, and each change can be accepted or rejected individually or reviewed as one set — making the operation inspectable, correctable and repeatable rather than a suggestion to paste somewhere else.

Key capabilities

  • Markdown authoring with live preview, links, backlinks, callouts, code blocks, math, diagrams, daily notes, search, and export.
  • Structured data stored in .kitable files, with typed records, formulas, filters, groups, views, attachments, and AI fields.
  • An infinite Whiteboard for diagrams, mind maps, flows, and freeform sketches, where agent proposals are reviewed before being applied.
  • A tool-using agent that can search the web, inspect documents and table schemas, update content, and save useful output into the active project.
  • Per-change review of agent document edits, with individual accept or reject controls and a combined diff view.
  • Trigger-and-action workflows that can be assembled visually, with testable steps, run history, and checks for missing inputs before a process is enabled.
  • Vision-powered table fields that extract structured JSON and plain OCR text from images into the same record.

Who uses it and how

  • Operations or marketing teams building campaign assets in batches: a key message and a face photo feed typed fields, and AI fields generate linked 16:9 and 9:16 thumbnail variants for every record, with the table acting as both queue and review surface.
  • Finance or administrative workflows that turn receipt photographs into searchable records, extracting vendor, address, category, structured JSON and OCR text into a single row for filtering, correction, summarisation, or hand-off to a review workflow.
  • Product and design teams expanding one product brief into a full asset pipeline — design candidates, orthographic views, a feature image, a lifestyle shot, a style board and launch copy — with each result attached to its source record.
  • Individuals and small teams who want a local-first knowledge base where agent output lands in their own files, and who are willing to review diffs before accepting them.

Getting started

Kition is distributed as a desktop application: the README directs users to download it from the project website at kition.ai or from the GitHub Releases page. The facts do not name a package-manager install, Docker image, or compose file.

How it compares

The repository metadata tags Kition as a notion-alternative, which places it alongside hosted, proprietary knowledge-base workspaces in the same broad category. Beyond that tag, the facts provided do not set out a feature-by-feature comparison with those products, so the contrast drawn here is limited to what Kition's own description supports: an Apache-2.0 licence, a local-first desktop model, and project files that stay on the user's machine.

When to use it — and when not to

Kition is explicitly in beta, and the README asks users to back up important workspaces and to review agent changes before relying on them in production workflows, so anyone who needs a settled system of record should wait or keep a separate copy of their data. The self-hoster takes on none of the usual server duties described here — no database, object storage or SMTP is mentioned, because this is a desktop application rather than a hosted service — but they do take on the responsibility of inspecting agent diffs and maintaining their own backups. Teams that want unreviewed, fully automatic AI output written straight into shared documents are a poor fit for a design that insists on the final say.

project readme (upstream, from github) — read inline

Kition

English · 简体中文 · Русский · 日本語 · Tiếng Việt · Français · Deutsch · Español

Documents, Tables, Whiteboards, Agents, and Workflows in one desktop workspace.
Write connected knowledge, think visually, build data tools, research the web, and automate repeatable work.

CI status Latest release Total GitHub release downloads License: Apache 2.0 Project status: beta macOS supported Windows supported Linux experimental

Download Kition

Website · Releases · Discussions · Contributing · Support · Security

Kition brings Markdown documents, structured table files, an infinite Whiteboard, a tool-using AI agent, web research, and visual workflows into one desktop workspace. Instead of making every task begin in a blank chat, Kition gives the Agent editable project files, typed records, attachments, and visible processes to work with. This makes powerful AI operations easier to inspect, correct, and repeat without handing context between disconnected tools.

Kition is currently in beta. Back up important workspaces and review agent changes before relying on them in production workflows.

Why Kition

  • Connected documents. Write Markdown with live preview, links, backlinks, callouts, code, math, diagrams, daily notes, search, and export.
  • Structured data beside the knowledge. Turn research and content into typed records, formulas, filters, groups, views, attachments, and AI fields.
  • Visual thinking with reviewable AI changes. Build diagrams, mind maps, flows, and freeform sketches, then review Agent proposals before applying them to the Whiteboard.
  • An Agent that can act. Search the web, inspect documents and table schemas, update content, and save useful output into the active project.
  • Reviewable document edits. Let the Agent edit the active document, inspect every addition and deletion, then accept or reject individual changes.
  • Visible automation. Assemble trigger-and-action workflows, test steps, inspect run history, and resolve missing inputs before enabling a process.

Let the Agent Edit—Keep the Final Say

Instead of returning another suggestion to copy and paste, the Kition Agent can read the active Markdown document, make scoped changes, and write the result back into the workspace. The document and the complete task trace remain visible together while the Agent works.

When a file changes outside the editor, Kition opens a document review surface that highlights additions, deletions, and rewrites. Each change can be accepted or rejected individually, or the complete edit can be reviewed as one set.

This creates a controlled document workflow: describe the goal in natural language, let the Agent edit the real file, inspect the diff, and decide exactly what remains in the final document.

Start With Work, Not a Blank Prompt

A capable Agent is useful, but a blank prompt still asks the user to design the task, describe every input, choose an output format, and recover when a step goes wrong. Kition moves that complexity into familiar work surfaces: documents, table fields, records, templates, and workflows.

The built-in scenarios below are ordinary .kitable files. Their prompts, field relationships, generated assets, and review states stay visible and can be adapted to a real project.

Generate campaign assets in batches

Start with a key message and a face photo. Typed fields keep the content type and intended emotion explicit, while AI fields generate linked 16:9 and 9:16 thumbnail variants for every record. The table becomes both the batch queue and the review surface.

Turn receipt images into searchable records

Drop receipt photos into an attachment field. Vision-powered fields extract the vendor, address, category, structured JSON, and plain OCR text directly into the same row. The results can then be filtered, corrected, summarized by the Agent, or passed into a review workflow.

Expand one product brief into a complete asset pipeline

One product concept can drive multiple design candidates, orthographic views, a feature image, a lifestyle shot, a style board, and launch copy. Each result remains attached to the source record, so the full chain is easier to compare, regenerate, and hand off than a collection of disconnected chat outputs.

Design posters from records, Boards, and the Agent

Design Studio turns a table record, a Board frame, or a generated image into an editable poster. Templates bind the workspace brand kit, size variants keep one composition across poster, story, square, and landscape, and the Agent proposes text and layout changes as a reviewable preview. Exports go back into documents and tables. See docs/design-studio.md.

These scenarios show the intended Kition loop: use documents for narrative context, tables for structured state, Whiteboards for visual thinking, the Agent for uncertain work, and Workflows for steps that should become repeatable.

Product Tour

Write connected documents

Write Markdown with live preview, internal links, backlinks, tags, outlines, embeds, callouts, code, math, diagrams, templates, daily notes, tabs, and full-text search. Export finished work to PDF when it needs to leave the workspace.

Turn data into useful tools

Create structured table files with typed fields, attachments, formulas, filters, sorting, grouping, multiple views, and fast record editing. Use them for research, content pipelines, lightweight CRM, inventory, or project tracking.

Give the agent real context

Ask the Kition agent to read and update documents, inspect table schemas, create or modify records, search the web, use tools, and follow a plan. Agent output can be written back into the active workspace.

Think and plan visually

Build diagrams, mind maps, flowcharts, sticky-note clusters, and freeform sketches on an infinite canvas. The Agent can inspect the board, propose typed changes, and let you accept or reject the preview before it becomes part of the workspace.

Automate repeatable work

Build trigger-and-action workflows on a visual canvas. Start from a template, assemble a workflow manually, test individual steps, inspect run history, and resolve missing inputs before enabling it.

Also included:

  • AI-assisted table fields for generating text, structured values, and media from record context.
  • Email connections and inbox-to-table workflows. See Email Connections.
  • Model, proxy, MCP, account, usage, update, and desktop integration settings.
  • A seeded Getting Started workspace with practical document, table, and workflow examples.

Install

Desktop builds are published through GitHub Releases.

  • macOS: download the latest .dmg.
  • Windows: download the latest installer.
  • Linux (experimental): download the latest x86_64 .AppImage.
  • Previous versions: browse the complete release history.

Run From Source

Requirements:

pnpm install --frozen-lockfile
pnpm dev

pnpm dev verifies and caches the pinned released runtime outside the Git worktree, then starts the Tauri desktop client.

For renderer, component, editor, and style work that does not require runtime integration:

pnpm dev:web

The web preview runs without the private runtime and is the fastest path for most client-side contributions. More options are documented in Runtime development.

Open Source Boundary

This repository contains the public React client and its Tauri desktop shell:

  • document, table, Whiteboard, Agent, settings, email, scenario, and workflow UI;
  • the Tauri desktop shell (src-tauri/) and desktop integration;
  • public runtime contracts, mocks, fixtures, and black-box tests;
  • packaging and unified release workflows.

The Kition runtime is maintained separately and its source is not included in this repository. Released desktop builds pair the public client with a signed runtime. Source checkouts communicate with it only through the public contracts in contracts/runtime/.

There is no Go source, Go module, or runtime source-build fallback in this repository.

Technology

Area Stack
Desktop Tauri 2 (Rust)
Interface React, TypeScript, Vite
Documents CodeMirror, Marked, Mermaid, KaTeX
Data and state IndexedDB, Jotai, Zod
Testing Vitest, Playwright

Contributing

Issues and pull requests for the public client are welcome.

  1. Read CONTRIBUTING.md for setup, scope, and required checks.
  2. Follow the Kition Development Standard, including the recommended Orca worktree workflow.
  3. Keep changes inside the public client/runtime contract boundary.
  4. Report security-sensitive issues through .github/SECURITY.md, not a public issue.

For product questions, account help, and private workspace concerns, see .github/SUPPORT.md.

Repository Layout

.
├── .github/                 # Community health files and CI workflows
├── contracts/runtime/       # Public runtime schemas
├── docs/                    # Product, contributor, design, and runtime docs
├── e2e/                     # Playwright black-box tests
├── public/                  # Static assets and onboarding files
├── scripts/                 # Repository checks and inspection entry points
├── src/                     # React application organized by feature domain
├── src-tauri/               # Tauri desktop shell, runtime manager, and packaging
├── tooling/                 # Build, lint, and test configuration
└── package.json             # Development, test, and packaging commands

License

Kition, including its runtime in crates/, is licensed under the Apache License 2.0.

Frequently asked questions

Is kition free to use?

kition is open source under the Apache-2.0 licence. There is no licence fee and no seat count — you can self-host it or, where the project offers one, pay a vendor for a managed version instead.

What does kition do?

Kition brings Markdown, DataTable, WhiteBoard, a tool-using AI agent, browser research, and visual workflows into one desktop workspace.

What is kition written in?

kition is primarily written in TypeScript. Its source is publicly available at https://github.com/KitionAI/kition, and it has 520 GitHub stars.