Show HN: Kote – Capture and reuse engineering context from AI chats and Git

Notifications You must be signed in to change notification settings

Fork 1

Star 3

BranchesTags

Open more actions menu

Folders and files

NameName

Last commit message

Last commit date

Latest commit

History

1,389 Commits

1,389 Commits

.github

.github

backend

backend

cli

cli

docs

docs

e2e

e2e

frontend

frontend

ide

ide

scripts

scripts

.codex

.codex

.dockerignore

.dockerignore

.env.example

.env.example

.gitignore

.gitignore

AGENTS.md

AGENTS.md

LICENSE

LICENSE

README.md

README.md

docker-compose.prod.yml

docker-compose.prod.yml

docker-compose.yml

docker-compose.yml

eslint.config.js

eslint.config.js

package-lock.json

package-lock.json

package.json

package.json

playwright.config.ts

playwright.config.ts

Repository files navigation

Stop losing your development context. Kote solves the everyday problems developers face when trying to recall past decisions, solutions, and technical rationale:

Never forget why you made that architectural change - Kote captures the reasoning behind commits, not just the code diff

Instantly recall how you solved that bug 6 months ago - Kote automatically surfaces relevant context from your AI sessions and Git history as you work

See relevant context right in your IDE - When you open a file, Kote displays accumulated knowledge and decisions directly in your editor

Access your development memory from anywhere - View accumulated knowledge in VS Code, CLI, WhatsApp, or the web interface

Unlike traditional documentation tools that require manual effort, Kote works passively in the background, capturing context as you work with AI assistants and Git.

Overview

Software development generates a high volume of technical context that rarely makes it into formal documentation:

Complex problem-solving discussions held with AI coding assistants (Claude Code, Copilot, ChatGPT).

Rationale behind architectural changes, which is often omitted in brief commit messages.

Project details, business rules, and domain-specific context.

Kote is designed to passively capture these workflows, organizing them into a unified knowledge base that can be used to ask questions, retrieve past decisions, and explore technical context when needed. Every time you open a relevant file in your IDE, Kote automatically retrieves the accumulated knowledge about it to help you understand its purpose, timeline of decisions, and the complete history behind it. All knowledge and AI analysis can be viewed in the Kote interface.

Demos

Kote Overview

Kote: capturing development context and querying knowledge.

CodeLens Integration

VS Code CodeLens showing related notes at the top of files.

How It Works

Kote integrates with your existing tools to collect and index technical knowledge automatically:

Development & AI Sessions: The VS Code extension automatically logs local AI conversations and saves highlighted code snippets to your knowledge base.

Git Workflow: The GitHub integration analyzes commit diffs on push, generating technical summaries and flagging potential operational issues.

Quick Notes: Send text or audio messages to Kote's messaging integrations to log quick notes or environment configurations.

CodeLens Integration: When opening files in VS Code, a CodeLens indicator appears at the top of files that have associated notes or decisions in your Kote. Click the indicator to view an AI-powered summary with timeline of key changes, understanding of the file context, and access to detailed notes.

Query & Retrieval: Query the accumulated knowledge base using natural language directly from the Web Application, the VS Code sidebar, WhatsApp or via CLI to ask questions or locate solutions, documents, and files.

Getting Started

To start using Kote:

Sign In: Go to pedro-duarte.ddns.net/kote and authenticate your account.

Configure Integrations: Connect your repository hosting (GitHub) via the Integrations dashboard in the web application.

Install the VS Code Extension: Install Kote from the VS Code Marketplace and sign in. Kote passively captures AI sessions in the background and lets you query your knowledge base using AI directly from VS Code.

Start Capturing Context Automatically: Once set up, Kote runs in the background and continuously captures AI sessions (Antigravity, Codex, Claude Code, OpenCode, etc.), GitHub activity (pushes and pull requests), and development context. Everything you work on becomes searchable knowledge that you can query to understand what you've done, why decisions were made, and how your system evolved over time.

Access Related Notes in Your IDE: When you open files in VS Code that have associated notes or decisions in your Kote, a CodeLens indicator appears at the top of the file showing the count of related notes. Click the indicator to view an AI-powered summary with timeline of key changes, understanding of the file context, and access to detailed notes with professional webview visualization, making it easy to access relevant context and decisions without leaving your development environment.

Tip

CodeLens not working? Make sure CodeLens is enabled in your VS Code settings ("editor.codeLens": true). It's enabled by default, but may have been disabled globally.

Self-Hosting (Docker)

If you prefer to run Kote on your own infrastructure or local machine, you can launch the entire stack using Docker Compose:

Clone the Repository:

git clone https://github.com/pedroaugusto04/knowledge-base.git cd knowledge-base

Configure Environment Variables: Copy the example environment file:

cp .env.example .env

Open the .env file and configure the essential keys to enable core features:

Admin Credentials: Change KB_ADMIN_EMAIL and KB_ADMIN_PASSWORD (used for your initial login).

AI Integrations (Search/Chat/Voice): Set KB_AUDIO_AI_API_KEY and KB_EMBEDDING_AI_API_KEY (Gemini API key is the default). Alternatively, you can self-host embeddings using the local Ollama service (e.g. nomic-embed-text-v2-moe) by configuring the KB_EMBEDDING_AI_PROVIDER=ollama environment variables.

File Storage: Fill in SUPABASE_URL, SUPABASE_SERVICE_ROLE_KEY, and KB_SUPABASE_STORAGE_BUCKET to store notes and attachments.

Start Services: Launch the database, message broker, backend API, and web application (database migrations will run automatically on startup):

docker compose up -d

Once running, access the local services:

Web Application: http://localhost:4311

API Server: http://localhost:4310

Tip

Point your VS Code Extension (knowledgeVault.apiUrl) or CLI (apiUrl in ~/.kb-config.json) to your self-hosted API URL (http://localhost:4310) to connect your editor and terminal to your local instance.

Features

Web Application & Knowledge Map

Interfaces to manage, search, and visualize captured knowledge.

Semantic chat interface for querying indexed data.

Detailed view of a captured note with metadata and tags.

Visual node graph illustrating relations between projects and notes.

VS Code Extension

Integrates directly with your editor to capture context during development.

Integrated sidebar containing AI chat and quick-save options.

Key Features:

CodeLens Integration: Automatically displays relevant notes and decisions at the top of files that have associated knowledge in your Kote. Click the CodeLens indicator to view an AI-powered summary with timeline of key changes, understanding of the file context, and access to detailed notes with professional webview visualization.

AI Chat Sidebar: Interactive AI chat interface for querying your knowledge base without leaving the editor.

Quick Save Commands: Save code selections or entire files as notes with right-click context menus.

AI Session History: View and search recent AI-assisted development sessions from various tools.

Real-time Sync: Monitor and sync local AI CLI sessions automatically in the background.

For configuration details, see ide/vscode/README.md.

GitHub Integration

Processes repository activity passively to record code changes and retrieve context.

Diff Analysis: Summarizes changes on every push.

Onboarding Backfill: After linking a repository, users can opt in to import recent commits as review notes. Configure the offer size with KB_GITHUB_BACKFILL_LIMIT (default: 5). Each imported commit consumes AI credits like a normal push review.

Alert System: Notifies the team via WhatsApp or Telegram if potential configuration or environmental issues are detected in a diff.

PR Context AI: Analyzes changed files and title/description of newly opened Pull Requests to automatically retrieve historical technical decisions and context, posting it as a PR comment.

CLI Tool (kote)

Synchronize terminal session histories and import local directories or files.

Importing AI session history from the terminal.

For installation steps and command options, see cli/README.md.

Browser Extension

Save documentation, issues, and articles directly from the web browser.

Browser extension popup for saving web content.

For setup instructions, see ide/browser-extension/README.md.

Model Context Protocol (MCP) Server

Provides developer memory retrieval and persistence directly to AI assistants (such as Cursor, Claude Desktop, Cline, and Antigravity).

Key Features:

kote_search_notes: Search developer notes, design decisions, and PR summaries using hybrid keyword and vector search, returning optimized snippets.

kote_get_note: Fetch the full Markdown body of a specific note by ID.

kote_create_note: Persistently save important development decisions or notes straight into your Kote memory graph.

Integration Quick Start: You can run the Kote MCP Server directly from npm using npx:

{ "mcpServers": { "kote": { "command": "npx", "args": ["-y", "@pedroaugusto04/kote-mcp"] } } }

For advanced configuration and setup, see ide/mcp/README.md.

Messaging Integrations (WhatsApp & Telegram)

Provides channels for logging quick notes and querying the database.

Configuration dashboard for WhatsApp, Telegram, and GitHub integrations.

Audio Notes: Transcribes and structures voice recordings into Markdown notes.

Image Capture: Upload screenshots or whiteboard diagrams to attach to projects.

Interactive Querying: Search the knowledge base using the /ask command.

License

See LICENSE for terms of use.

About

A developer memory layer that automatically captures and organizes AI sessions, Git history, and development context into searchable knowledge.

pedro-duarte.ddns.net/kote/

Topics

productivity

documentation

ai

self-hosted

developer-tools

knowledge-base

semantic-search

knowledge-management

ai-agents

context-engineering

Resources

Readme

License

View license

Uh oh!

There was an error while loading. Please reload this page.

Activity

Stars

3 stars

Watchers

0 watching

Forks

1 fork

Report repository

Releases 1

Kote v1.0.0

Latest

Jul 1, 2026

Packages 0

Uh oh!

There was an error while loading. Please reload this page.

Contributors

Uh oh!

There was an error while loading. Please reload this page.

Languages

TypeScript 75.7%

JavaScript 15.5%

CSS 6.1%

HTML 2.3%

Shell 0.2%

Handlebars 0.2%