AI-Powered Agent Framework

Meet Wisemonkey
Your Intelligent Agent

A modular, multi-provider AI agent with persistent memory, vector search, MCP support, and a powerful tool system, all in a sleek terminal interface.

See It In Action

A simulated Wisemonkey session — the welcome screen you see on first launch, followed by example prompts showcasing natural-language control over tools, memory, providers, and MCP servers.

wisemonkey — default
.-"-. _/.-.-.\_ ( ( o o ) ) |/ " \| \ ݁݁ ⏝ / /`"""`\ / \ ██ ██ ██ ▄█████ ██████ ██▄ ▄██ ▄████▄ ███ ██ ██ ▄█▀ ██████ ██ ██ ██ ▄█▄ ██ ██ ▀▀▀▄▄▄ ██▄▄ ██ ▀▀ ██ ██ ██ ██ ▀▄██ ████ ██▄▄ ▀██▀ ▀██▀██▀ ██ █████▀ ██▄▄▄▄ ██ ██ ▀████▀ ██ ██ ██ ▀█▄ ██▄▄▄▄ ██
Wisemonkey v2026.6.2 commit: 915a193
Up to date

Session restored: 'wisemonkey'
location:      ~/.local/share/wisemonkey/sessions/wisemonkey
working dir:  ~/Projects/wisemonkey
created:      2026-05-29 08:27:51
last accessed: 2026-06-05 11:31:08

Built for Power Users

🧠

Multi-Provider Support

Connect to OpenAI, Anthropic, Ollama, LM Studio, or any OpenAI-compatible API. Switch providers on the fly.

💾

Persistent Memory

Session-scoped chat history, user profiles, and notes that survive restarts. Your agent remembers everything.

🔍

Vector Search (RAG)

Embed PDFs and documents into a ChromaDB vector store. Retrieve relevant context on demand.

🔌

MCP Integration

Connect to Model Context Protocol servers for filesystem access, GitHub, databases, and more.

🛠️

Dynamic Tool Registry

Register tools at runtime. MCP server tools are auto-discovered and injected into the agent's capabilities.

Streaming & Reasoning

Real-time token streaming with configurable reasoning effort. Watch your agent think step by step.

Up and Running in Seconds

$ curl -fsSL https://codeberg.org/langurmonkey/wisemonkey/raw/branch/master/install.sh | bash $ wisemonkey
$ git clone https://github.com/wisemonkey/wisemonkey.git
$ cd wisemonkey
$ uv run wisemonkey

One File, Total Control

On first run, Wisemonkey creates ~/.config/wisemonkey/config.yaml with sensible defaults. Works with any OpenAI-compatible endpoint out of the box.

model:
  # Model name
  name: qwen/qwen3.6-35b-a3b
  base_url: http://127.0.0.1:1234/v1
  temperature: 0.8
  reasoning_effort: medium
  reasoning_visible: false
 
embedding:
  name: qwen/qwen3-embedding-0.6b-gguf
  base_url: http://127.0.0.1:1234/v1
 
agent:
  max_turns: 50
  markdown: false
  max_chat_history: 128000
  vi_mode: false
{
  "mcpServers": {
    "filesystem": {
      "command": "npx",
      "args": ["-y", "@modelcontextprotocol/server-filesystem", "/tmp"]
    }
  }
}
 
  # Use /mcp edit to modify, /mcp tools to list
# API key — loaded automatically by python-dotenv
OPENAI_API_KEY=your-api-key-here
 
# Searched in order:
# ./.env → ~/.config/wisemonkey/.env → ~/.env
🔌

Any OpenAI-Compatible Endpoint

LM Studio, Ollama, OpenWebUI, or your own server. Just set base_url.

🧠

Reasoning Control

Set reasoning_effort to none, low, medium, or high. Toggle visibility with reasoning_visible.

📐

Embeddings

Separate embedding model config. Use any OpenAI-compatible embedding endpoint for vector search.

⌨️

vi Mode

Enable vi_mode for vim-style editing. Press v in normal mode to open your external editor.

Wisemonkey at a Glance

5+
AI Providers
MCP Servers
14
Built-In Tools
1
Binary
0
Config Headaches