freshcrate
Skin:/
Home > MCP Servers > kernel

kernel

kbot — the AI agent that dreams, learns, and evolves. 764+ tools, 35 agents, 20 providers. Music production, iPhone control, financial analysis, cyber threat intel. Always-on daemon. Runs offline. npm

Why this rank:Release freshnessStrong adoptionHealthy release cadence

Description

kbot — the AI agent that dreams, learns, and evolves. 764+ tools, 35 agents, 20 providers. Music production, iPhone control, financial analysis, cyber threat intel. Always-on daemon. Runs offline. npm i -g @kernel.chat/kbot

README

kbot
Open-source terminal AI agent. 35 agents. 600+ tools. 20 providers. Science, finance, security, and more.

kbot demo

CI npm version npm downloads MIT License kernel.chat Discord kbot MCP server

If kbot is useful to you, consider starring the repo — it helps others discover the project.

npm install -g @kernel.chat/kbot

Why kbot?

Most terminal AI agents lock you into one provider, one model, one way of working. kbot doesn't.

  • 20 providers, zero lock-in — Claude, GPT, Gemini, Grok, DeepSeek, Groq, Mistral, SambaNova, Cerebras, OpenRouter, and more. Switch with one command.
  • Runs fully offline — Embedded llama.cpp, Ollama, LM Studio, or Jan. $0, fully private.
  • Learns your patterns — Bayesian skill ratings + pattern extraction. Gets faster over time.
  • 35 specialist agents — auto-routes your request to the right expert (coder, researcher, writer, guardian, quant, and 30 more).
  • 600+ tools — files, bash, git, GitHub, web search, deploy, database, game dev, VFX, research, science, finance, security, and more.
  • Programmatic SDK — use kbot as a library in your own apps.
  • MCP server built in — plug kbot into Claude Code, Cursor, VS Code, Zed, or Neovim as a tool provider.

How it compares

kbot Claude Code Codex CLI Aider OpenCode
AI providers 20 1 1 6 75+
Specialist agents 35 0 0 0 0
Built-in tools 600+ ~20 ~15 ~10 ~15
Science tools 114 0 0 0 0
Learning engine Yes No No No No
Offline mode Embedded + Ollama No No Ollama Ollama
SDK Yes No Yes No No
MCP server Yes N/A No No No
Web companion kernel.chat No No No No
Open source MIT Source available Apache 2.0 Apache 2.0 MIT
Cost BYOK / $0 local $20+/mo BYOK BYOK BYOK

Quick Start

# Install globally
npm install -g @kernel.chat/kbot

# Or run directly (no install)
npx @kernel.chat/kbot

# Or use the install script (auto-installs Node.js if needed)
curl -fsSL https://kernel.chat/install.sh | bash
# Interactive mode
kbot

# One-shot
kbot "explain this codebase"
kbot "fix the auth bug in src/auth.ts"
kbot "create a Dockerfile for this project"

# Pipe mode (for scripting)
kbot -p "generate a migration for user roles" > migration.sql

# Use local models (free, no API key)
kbot local

# Set up your API key
kbot auth

Specialists

kbot auto-routes to the right agent for each task. Or pick one with --agent <name>.

Agents
Core kernel, researcher, coder, writer, analyst
Extended aesthete, guardian, curator, strategist
Domain infrastructure, quant, investigator, oracle, chronist, sage, communicator, adapter
Presets claude-code, cursor, copilot, creative, developer

600+ Tools

Category Examples
Files & Code read, write, glob, grep, lint, format, type-check
Shell bash, parallel execute, background tasks
Git & GitHub commit, diff, PR, issues, code search
Web search, fetch, browser automation, browser agent
Research arXiv, Semantic Scholar, HuggingFace, NASA, DOI
Data CSV read/query/write, transforms, reports, invoices
Quality lint (ESLint/Biome/Clippy), test (Vitest/Jest/pytest), deps audit, formatting
Deploy Vercel, Netlify, Cloudflare Workers/Pages, Fly.io, Railway
Database Postgres, MySQL, SQLite queries, Prisma, ER diagrams, seed data
Containers Docker build/run/compose, Terraform
Creative p5.js generative art, GLSL shaders, SVG patterns, design variants
VFX GLSL shaders, FFmpeg, ImageMagick, Blender, procedural textures
Game Dev 16 tools for Godot, Unity, Unreal, Bevy, Phaser, Three.js, PlayCanvas, Defold
Training dataset prep, fine-tuning, evaluation, model export
Social post to X, LinkedIn, Bluesky, Mastodon — single posts and threads
Sandbox Docker sandboxes, E2B cloud sandboxes, isolated code execution
Notebooks Jupyter read/edit/insert/delete cells
Build Matrix cross-platform builds — mobile, desktop, WASM, embedded, server
LSP goto definition, find references, hover, rename, diagnostics
Memory persistent memory save/search/update across sessions
MCP marketplace search/install, 20 bundled servers
IDE MCP server, ACP server, LSP bridge
Forge create tools at runtime, publish to registry, install from registry
Meta subagents, worktrees, planner, sessions, checkpoints, self-eval
Science & Math symbolic compute, matrix ops, FFT, ODEs, probability, optimization, graph theory, OEIS
Physics orbital mechanics, circuits, signal processing, particles (PDG), relativity, quantum simulator, beam analysis, fluid dynamics
Chemistry PubChem compounds, reactions, periodic table (118 elements), spectroscopy, stoichiometry, thermodynamics
Biology PubMed, gene lookup, protein/PDB, BLAST, drug/ChEMBL, pathways, taxonomy, clinical trials
Earth & Climate earthquakes/USGS, climate/NOAA, satellite imagery, geology, ocean, air quality, volcanoes, water resources
Neuroscience brain atlas, EEG analysis, cognitive models, neural simulation, connectome, psychophysics
Social Science psychometrics, game theory, econometrics, social network analysis, survey design, voting systems
Humanities corpus analysis, formal logic, argument mapping, ethics frameworks, historical timelines, stylometry
Health & Epidemiology SIR/SEIR models, epidemiology calculations, disease surveillance, nutrition, vaccination modeling
Finance market data, technical analysis, paper trading, DeFi yields, wallet & swaps, stock screener, sentiment
Cybersecurity dep_audit, secret_scan, ssl_check, headers_check, cve_lookup, port_scan, owasp_check
Self-Defense memory HMAC, prompt injection detection, knowledge sanitization, forge verification, anomaly detection

20 Providers

Provider Cost Local?
Embedded (llama.cpp) Free Yes
Ollama Free Yes
LM Studio Free Yes
Jan Free Yes
DeepSeek $0.27/M in Cloud
SambaNova $0.50/M in Cloud
Cerebras $0.60/M in Cloud
Groq $0.59/M in Cloud
NVIDIA NIM $0.80/M in Cloud
Together AI $0.88/M in Cloud
Fireworks AI $0.90/M in Cloud
Google (Gemini) $1.25/M in Cloud
Mistral $2.00/M in Cloud
OpenAI (GPT) $2.00/M in Cloud
Cohere $2.50/M in Cloud
Anthropic (Claude) $3.00/M in Cloud
xAI (Grok) $3.00/M in Cloud
Perplexity $3.00/M in Cloud
OpenRouter varies Cloud
kbot local Free Yes

Set any provider's env var and kbot auto-detects it. Or run kbot auth for interactive setup.

SDK

import { agent, tools, providers } from '@kernel.chat/kbot'

const result = await agent.run("fix the auth bug", { agent: 'coder' })
console.log(result.content)

for await (const event of agent.stream("explain this code")) {
  if (event.type === 'content_delta') process.stdout.write(event.text)
}

Architecture

graph TD
    A[User Message] --> B{Local-first check}
    B -->|file, git, grep| C[Instant response — $0]
    B -->|Needs AI| D{Complexity detection}
    D -->|Simple| E[Bayesian Router → Specialist]
    D -->|Multi-step| F[Autonomous Planner]
    E --> G[Provider API + Tool Loop]
    F --> G
    G --> H{600+ Tools}
    H --> I[File ops, bash, git, GitHub, search, deploy, DB, game dev...]
    G --> J[Learning Engine]
    J --> K[Patterns + Solutions + User Profile]
    G --> L[Checkpointing]
    L --> M[Resume from last tool call]
Loading

MCP Server

kbot MCP server

Use kbot as a tool provider inside any MCP-compatible IDE:

{
  "mcp": {
    "servers": {
      "kbot": { "command": "kbot", "args": ["ide", "mcp"] }
    }
  }
}

Or run via Docker:

docker run -i --rm kernelchat/kbot ide mcp

Works with Claude Code, Cursor, VS Code, Windsurf, Zed, Neovim.

Commands

Command What it does
kbot Interactive REPL
kbot "prompt" One-shot execution
kbot auth Configure API key
kbot local Use local AI (Ollama, embedded, LM Studio, Jan)
kbot serve Start HTTP REST + SSE streaming server
kbot audit <repo> Security + quality audit of any GitHub repo
kbot contribute <repo> Find good-first-issues and quick wins
kbot share Share conversation as GitHub Gist
kbot pair File watcher with auto-analysis
kbot team Multi-agent TCP collaboration
kbot record Terminal session recording (SVG, GIF, asciicast)
kbot voice Text-to-speech output mode
kbot watch Real-time file analysis on change
kbot bootstrap Outer-loop project optimizer (visibility scoring)
kbot plugins Search, install, update community plugins
kbot models List, pull, remove, catalog local models
kbot changelog Generate changelog from git history
kbot completions Shell completions (bash, zsh, fish)
kbot cloud Sync learning data to kernel.chat
kbot ide mcp Start MCP server for IDEs
kbot doctor 10-point health check

Power-User Flags

kbot --architect "design the auth system"    # Architecture mode — plan before code
kbot --thinking "solve this hard problem"    # Extended reasoning with thinking budget
kbot --self-eval "write a parser"            # Self-evaluation loop — scores and retries
kbot --computer-use "fill out this form"     # Computer use — controls mouse and keyboard
kbot -p "query" > output.txt                 # Pipe mode — clean output for scripting

Security

  • API keys encrypted at rest (AES-256-CBC)
  • Destructive operations require confirmation
  • Shell commands sandboxed with blocklist
  • Tool execution timeout (5 min) with middleware pipeline
  • Config files restricted to owner (chmod 600)
  • 0 P0/P1 security issues (audited March 2026)

Development

cd packages/kbot
npm install
npm run dev          # Run in dev mode
npm run build        # Compile TypeScript
npm run test         # Run tests (vitest)

Web Companion — kernel.chat

kbot has a web companion at kernel.chat — same agents, persistent memory, and a visual interface. Free to use (20 messages/day).

Community

License

MIT — kernel.chat group

Release History

VersionChangesUrgencyDate
v3.97.0## kbot v3.97.0 — Stream V2 + Intelligence Coordinator 23 releases since v3.74.0. This is a major milestone covering the Stream V2 platform, a full cognitive architecture, and self-hosting capabilities. ### Stream V2 Platform - **PCM audio engine** with procedural sound generation - **Overlay system** — weather effects, chat overlays, viewer stats - **AI-powered chat** with 31 commands for viewer interaction - **VOD recording and playback** support - **Multi-platform streaming** — Twitch, KickHigh4/6/2026
v3.74.0## K:BOT v3.74.0 ### Quality & Security Improvements - MCP server version now dynamically matches package.json - Error sanitization across all MCP servers (prevents leaking internal paths, API keys, URLs) - TDQS-compliant tool descriptions (Purpose Clarity, Usage Guidelines, Behavioral Transparency, Parameter Semantics, Conciseness, Contextual Completeness) - Input validation constraints (min/max length, regex patterns) on all MCP tool parameters - Path traversal prevention on all file-handlingMedium4/4/2026
v3.59.0## What's New ### Desktop Agent (Computer-Use) - 866+ lines added to `computer.ts` — screenshot, click, type, scroll, drag, key combos - kbot can now drive any macOS GUI app (Ableton, Figma, browsers, etc.) - Native screenshot filtering via MCP server ### Session Isolation (Serve Mode) - `memory.ts` rewritten with `Map<sessionId, ConversationTurn[]>` — no more shared state - `serve.ts` creates unique session per HTTP request, destroys after response - Concurrent `/stream` requests fully isolatMedium3/31/2026
v3.58.0## The Hacker's Terminal 38 new security tools. 16,575 lines. kbot is now a complete offensive and defensive security platform. ### Security Brain Embedded knowledge base with full MITRE ATT&CK (14 tactics, 65+ techniques), OWASP Top 10 2025, Lockheed Martin Kill Chain, CVE pattern library, and tool mapping. - `security_brain` — query the knowledge base - `attack_lookup` — look up any ATT&CK technique by ID or keyword - `killchain_analyze` — map attacks to kill chain stages ### CTF Platform BMedium3/30/2026
v3.57.0## What's New ### New AI Models - **GPT-5.4** + GPT-5.4-pro — new OpenAI default model - **DeepSeek V4** — new DeepSeek default model - **Mistral Small 2503** — new Mistral fast model - **Voxtral TTS** — Mistral's text-to-speech (API + local binary) - **Gemini 3.1 Flash Lite** — added to Google models - **mistral-small3:22b** — added to Ollama local models ### New Features - **Visa Agent Payments** (`visa_status`, `visa_pay`, `visa_history`, `visa_refund`) — tokenized card payments for AI agenMedium3/30/2026
v3.56.0## What's New **6 new CLI command groups, 32 new tools** — everything manageable from terminal. ### New Commands | Command | What it does | |---------|-------------| | `kbot admin users` | List/search platform users | | `kbot admin billing mrr` | Check MRR via Stripe | | `kbot admin moderate` | Moderation queue | | `kbot monitor` | Platform health dashboard | | `kbot monitor --uptime` | Service health check | | `kbot monitor --alerts` | Active alerts | | `kbot deploy all` | Ship everything inMedium3/30/2026
v3.52.0## 10 New Features ### `kbot bench` — Self-Evaluation Run standardized benchmarks against 20 coding and research tasks across 6 categories (codegen, bugfix, refactor, explain, research, science). Scores pattern matching, tool usage, and speed. History tracking and run comparison. ### `kbot lab` — Interactive Science REPL Drop into a science-focused session with 10 domains (physics, chemistry, biology, math, neuroscience, earth, social, humanities, health, general). Auto-logs to research noteboMedium3/29/2026
v3.41.0## What's New (v3.33.0 → v3.41.0) ### v3.41.0 — Memory Hot-Swap, Side Conversations, Interactive Buttons - Memory hot-swap: switch memory contexts without restarting - Side conversations: branch off into parallel threads - Interactive buttons in terminal UI ### v3.40.0 — Shared Intelligence - Pattern feed: share learned patterns across kbot instances - Knowledge base: persistent collective knowledge store - User graph: relationship-aware context - Email delivery fixes ### v3.39.0 — AutonomousMedium3/26/2026
v3.32.0## The Kernel Stack This release introduces **The Kernel Stack** — a manifesto and architecture document for the Claude Code + kbot development system. ### What's in the Kernel Stack | Layer | What | |-------|------| | **Orchestrator** | Claude Code (Opus 4.6, 1M context) | | **Agent Framework** | kbot — 350+ tools, 26 specialists, learning engine | | **Protocol** | MCP — connects GitHub, Discord, Playwright, Supabase, local AI | | **Agents** | 36 parallel specialists (ship, QA, security, gitMedium3/24/2026
v3.31.1# kbot v3.31.1 — The Biggest Update Yet **12 versions shipped in one session.** v3.26.0 → v3.31.1. ## What's New ### v3.31 — Daemon Agent - **`kbot daemon start`** — background intelligence that never sleeps - 5 subsystems: market watch (15 min), security patrol (1 hr), synthesis (2 hr), health check (30 min), episodic digest (6 hr) - **`kbot briefing`** — morning intelligence report (market, security, stats, actions) - **Notification system** — macOS/Linux/Windows native notifications, DiscoMedium3/24/2026
v3.12.0## What's New ### Replit Integration kbot now auto-detects Replit environments and adapts: - **Lite mode** (`--lite`) — skips heavy tools (Docker, browser, local models, GPU) to fit cloud IDE resource limits - **Auto-detection** — Replit env vars (`REPL_ID`, `REPL_SLUG`) trigger lite mode automatically - **Replit Secrets** — API keys from Replit Secrets are auto-detected - **Public URL** — serve mode shows your Repl's public URL on startup ### Replit Template Fork-ready template at `templates/Low3/20/2026
v3.11.0## 8 Research Breakthroughs — No Other CLI Agent Has Any of These ### v3.10.0 1. **Voyager Skill Library** — successful tool chains auto-distill into reusable skills indexed by embedding 2. **MAR Multi-Persona Reflection** — 5 specialists critique failures, judge synthesizes lessons 3. **Spec-Driven Development** — \`kbot spec\` generates formal requirements before coding (EARS notation) 4. **Three-Tier Memory Synthesis** — observations → reflections → identity (Stanford Generative Agents) ###Low3/20/2026
v3.7.1## What's New ### Full Cognitive Stack (v3.7.0) All 11 cognitive modules now fire on every agent cycle: - Free Energy, Predictive Processing, Reasoning, Intentionality, Temporal, Confidence, Strange Loops, Integrated Information — plus the existing Learning, Self-Eval, and Skill Rating. ### Skills Auto-Discovery Drop `.md` files in `~/.kbot/skills/` or `.kbot/skills/` and kbot auto-loads them as context. The equivalent of Copilot Agent Skills. ### Autotelic Agent New `kbot autotelic` commLow3/19/2026
v3.7.0## The Intelligence Upgrade kbot had 11 cognitive modules. Only 3 were wired into the agent loop. The other 8 were dead code — architecturally beautiful TypeScript implementations of peer-reviewed research papers that never actually ran. **v3.7.0 wires them all in.** ### Now Active in Every Agent Cycle | Module | Paper | What It Does | |--------|-------|-------------| | Free Energy | Friston, 2010 | Explore vs exploit decisions | | Predictive Processing | Clark, 2013 | Anticipates user's nexLow3/19/2026
v3.6.1Documentation-only release. Both READMEs now surface 60% more features that existed in code but weren't documented: - **9 new tool categories**: Quality, Creative, Training, Sandbox, Notebooks, Build Matrix, LSP, Memory, Forge - **9 new CLI commands**: `voice`, `watch`, `bootstrap`, `plugins`, `models`, `changelog`, `completions`, `cloud`, `pair`/`team`/`record` - **5 power-user flags**: `--architect`, `--thinking`, `--self-eval`, `--computer-use`, `-p` No code changes. Install/upgrade: `npm iLow3/19/2026
v3.6.0## What's New in v3.6.0 ### Limitless Execution (5 new patterns) - **Tool Discovery** — smart error messages guide AI to `mcp_search → mcp_install → forge_tool` - **Cost-Aware Model Routing** — trivial tasks use fast model, complex tasks use full model - **Goal Decomposition** — `routeStepToAgent()` assigns specialist agents per plan step - **Fallback Chains** — URL fails → web search. Command not found → npx. Auto-recovery. - **Self-Extension** — `forge_tool` creates tools at runtime, sandboxeLow3/19/2026

Dependencies & License Audit

Loading dependencies...

Similar Packages

oh-my-piâŒĨ AI Coding agent for the terminal — hash-anchored edits, optimized tool harness, LSP, Python, browser, subagents, and morev15.9.5
tweetsave-mcp📝 Fetch Twitter/X content and convert it into blog posts using the MCP server for seamless integration and easy content management.main@2026-06-01
open-coworkOpen-source AI agent desktop app for Windows & macOS. One-click install Claude Code, MCP tools, and Skills — with sandbox isolation, multi-model support, and Feishu/Slack integration.v3.3.1
ori-cliAgentic coding harness with persistent memory and a REPL body. Built on Ori Mnemos. Open source must win.master@2026-05-08
polymarket-trader-mcpThe most comprehensive MCP server for Polymarket — 48 tools spanning direct trading, market discovery, smart money tracking, copy trading, backtesting, risk management, and portfolio optimization. Worv1.6.7

More in MCP Servers

AstrBotAgentic IM Chatbot infrastructure that integrates lots of IM platforms, LLMs, plugins and AI feature, and can be your openclaw alternative. ✨
agentscopeBuild and run agents you can see, understand and trust.
claude-plugins-officialOfficial, Anthropic-managed directory of high quality Claude Code Plugins.
langchain4jLangChain4j is an open-source Java library that simplifies the integration of LLMs into Java applications through a unified API, providing access to popular LLMs and vector databases. It makes impleme