dendrux 0.1.0a1

The runtime for agents that act in the real world.

dendrux

Python SDK for Dendrux — the framework for building agents with tools, persistence, and observability.

Version: 0.1.0a1

Install

pip install -e ".[all]"                # Everything
pip install -e ".[anthropic,db]"       # Just Anthropic + SQLite
pip install -e ".[openai,db]"          # Just OpenAI + SQLite

Quick Example

import asyncio
from dendrux import Agent, tool
from dendrux.llm.anthropic import AnthropicProvider

@tool()
async def add(a: int, b: int) -> int:
    """Add two numbers."""
    return a + b

async def main():
    async with Agent(
        provider=AnthropicProvider(model="claude-sonnet-4-6"),
        prompt="You are a calculator.",
        tools=[add],
    ) as agent:
        result = await agent.run("What is 15 + 27?")
        print(result.answer)

asyncio.run(main())

Providers

Provider	Import	Use case
Anthropic	from dendrux.llm.anthropic import AnthropicProvider	Claude models
OpenAI	from dendrux.llm.openai import OpenAIProvider	GPT models + vLLM, SGLang, Groq, Ollama
OpenAI Responses	from dendrux.llm.openai_responses import OpenAIResponsesProvider	GPT + built-in tools (web search)
Mock	from dendrux.llm.mock import MockLLM	Deterministic testing

API Quick Reference

from pathlib import Path
from dendrux.observers.console import ConsoleObserver

async with Agent(
    provider=provider,                  # Required: LLM provider
    prompt="...",                        # Required: system prompt
    tools=[add],                         # Optional: tool functions
    database_url=f"sqlite+aiosqlite:///{Path.home() / '.dendrux' / 'dendrux.db'}",
    redact=my_scrubber,                  # Optional: scrub persisted strings
) as agent:
    result = await agent.run(
        "What is 15 + 27?",
        observer=ConsoleObserver(),      # Optional: terminal output
        tenant_id="org-123",             # Optional: multi-tenant isolation
        metadata={"thread": "t1"},       # Optional: your linking data
    )

RunResult

result.answer          # str | None — the agent's final answer
result.status          # RunStatus — SUCCESS, ERROR, MAX_ITERATIONS, WAITING_CLIENT_TOOL, CANCELLED
result.steps           # list[AgentStep] — full reasoning chain
result.iteration_count # int — how many loop iterations ran
result.usage           # UsageStats — input_tokens, output_tokens, total_tokens
result.run_id          # str — unique run identifier (ULID)
result.error           # str | None — error message if status is ERROR

Tool Options

@tool()                                  # Basic server tool
@tool(target="client")                   # Client-side — agent pauses
@tool(max_calls_per_run=3)               # Limit calls per run
@tool(timeout_seconds=120)               # Custom timeout (default 120s)
@tool(parallel=False)                    # Run alone, not concurrently

Full Documentation

See the main README for provider setup, configuration, database guide, CLI, dashboard, observer system, and examples.