Overview

Quenda is a lightweight, layered Agent framework for Python. It provides a minimal API surface — Agent, Session, @tool — and ships with a terminal-based AI coding agent called Quenda Code. The framework is designed around a strict four-layer architecture (Kernel → Runtime → Host → Interface) where each layer depends only on the one inside it, making the core model-tool loop fully testable without network access.

Key Features

  • Minimal API — Agent, Session, @tool, and you’re done
  • 26 model providers — OpenAI, Anthropic, DeepSeek, DashScope, Moonshot, Ollama, and 20 more, all behind one registry
  • 9 core tools — Filesystem, shell, Python sandbox, and user interaction, all workspace-scoped
  • Skills framework — Composable capability packages with instructions, resources, and tools
  • Security by code — SSRF protection, command filtering, import restrictions, workspace isolation
  • Observable by default — Every run emits structured events for streaming and debugging
  • Context compression — Automatic summarization when context grows large

Quenda Code

The flagship application is Quenda Code, an AI coding agent that runs in your terminal:

pip install quenda quenda-code
quenda code

It reads your codebase, writes code, runs commands, and helps you ship — with code-aware understanding, workspace-scoped file operations, safe shell execution, session persistence, and interactive mode switching between code, architect, and chat.

Architecture

Interface → Host → Runtime → Kernel
Layer Responsibility
Kernel Synchronous model-tool loop. No knowledge of agents/sessions.
Runtime Async Agent/Session/Run lifecycle, event emission, context mgmt.
Host Persistence, identity, permissions, instruction composition.
Interface Event rendering, user interaction, REPL.

SDK Quick Start

from quenda import Agent, tool
from quenda.providers import get_provider_registry
from quenda.tools import get_core_tools
import asyncio

@tool
def calculate(expression: str) -> float:
    """Safely evaluate a math expression."""
    import ast
    node = ast.parse(expression, mode='eval')
    return eval(compile(node, '<string>', 'eval'), {"__builtins__": {}}, {})

model = get_provider_registry().get_model("deepseek", "deepseek-v4-flash")

agent = Agent(
    name="assistant",
    system_prompt="You are a helpful assistant.",
    tools=[calculate, *get_core_tools(".")],
    model=model,
)

async def main():
    session = agent.open_session()
    result = await session.send("What is 15% of 847?")
    print(result)

asyncio.run(main())

Installation

# Quenda Code — AI coding assistant (CLI)
pip install quenda quenda-code

# Quenda SDK — Build agents in Python
pip install quenda

Requires Python 3.12+. Zero required runtime dependencies.

License

MIT