Everything about local-first AI workflows
From install and agents to hooks, model picks, and concrete creator / developer / researcher workflows — we put answers to writing.
Browse by topic
12 categories
10 posts
Getting started
Install, first-run, migration, and core concepts — Kition from zero.
View category →12 posts
AI agent deep dives
Plan mode, hooks, subagents, MCP — get the agent to actually ship.
View category →10 posts
Data tables
Fields, views, AI field, CSV — structured data back inside your vault.
View category →10 posts
Documents & Markdown
Wikilinks, callouts, palette, templates — Obsidian-style writing.
View category →8 posts
Local-first & privacy
Where your data lives, why local, offline, and Git sync.
View category →8 posts
Models & providers
OpenAI, Anthropic, AiApiy, custom, China-hosted — how to pick.
View category →8 posts
Creators
Xiaohongshu, newsletters, YouTube, blogs — AI-accelerated creation.
View category →8 posts
Developers
Tech notes, PR review, RFCs, CI — agent meets engineering.
View category →7 posts
Researchers
Literature review, PDF RAG, Zotero, paper writing — academic flows.
View category →7 posts
Teams & consultants
Consulting, sales, PM, HR, lawyers, startups — operating workflows.
View category →8 posts
Compare & migrate
Deep dives: Kition vs Notion / Obsidian / Cursor / ChatGPT.
View category →4 posts
Advanced / Hooks / MCP
Hooks, MCP server configuration, custom tool development — power users.
View category →Latest posts
Most recent first
Building a custom agent tool: from schema to deployment
When built-ins aren’t enough, write your own. 10 lines + a JSON schema = one new agent capability.
Read more →Using hooks for audit and compliance
Law, healthcare, education need "every agent action traceable." Hooks + log hashing = a low-cost audit setup.
Read more →Complete guide to configuring MCP servers
stdio / SSE / HTTP transports; auth, timeouts, scopes — from beginner to production-grade.
Read more →Kition Hooks: events, filters, side-effects
Hooks are the agent middleware — slot code before/after tool calls to control behavior and emit audit trails.
Read more →Kition vs Logseq: open-source outliner vs commercial agent workspace
Both local-first. Logseq is an open-source outliner with a strong research community; Kition targets AI-era daily productivity.
Read more →Kition vs Roam Research: block refs vs Markdown + tables
Roam’s block-refs left a deep mark. Kition picks Markdown + tables — broader reach and easier migration.
Read more →Kition vs Reflect: two takes on linked notes + AI
Reflect: cloud linked notes + GPT enhancement. Kition: local + agent + tables. Different focuses.
Read more →Kition vs ChatGPT Desktop: chat history vs vault crystallization
ChatGPT history isn’t "work product." Kition lands AI output into a searchable, structured vault.
Read more →Kition vs Claude Desktop: agent + workspace vs agent only
Claude Desktop is chat-first; Kition is workspace-first. The former is great at chat; the latter ships artifacts.
Read more →Ready when you are.
Kition is a local-first AI workspace. Markdown documents, structured tables, and an AI agent — running on your own machine, against the model provider you choose.