aigie 0.2.44

Enterprise-grade AI agent reliability monitoring and autonomous remediation

Aigie Python SDK

Aigie is the official Python SDK for Kytte -- the Autonomous Reliability Platform for Agentic AI.

Proprietary software. Copyright (c) 2026 Kytte AI, Inc. All Rights Reserved. Use of the Aigie SDK requires a written agreement with Kytte AI, Inc. and a valid Kytte account. Not licensed for redistribution or modification. Contact: [e-mail]

Kytte is a runtime reliability platform for agentic AI, built to address the core reason most AI agents never reach production: non-deterministic runtime failures. Unlike reactive observability tools, Kytte operates during execution -- proactively detecting and autonomously remediating failed agent steps in real time, with no changes to your existing infrastructure or agent architecture. The Aigie SDK is the integration layer that connects your agents to the Kytte platform:

• Detects context drift and errors before they impact users
• Fixes issues automatically through self-healing workflows
• Prevents failures with predictive intervention

Installation

pip install aigie

With optional integrations

# Compression (recommended for production -- 50-90% bandwidth savings)
pip install aigie[compression]

# LLM providers
pip install aigie[openai]              # OpenAI
pip install aigie[anthropic]           # Anthropic Claude
pip install aigie[gemini]              # Google Gemini

# Agent frameworks
pip install aigie[langchain]           # LangChain
pip install aigie[langgraph]           # LangGraph
pip install aigie[claude-agent-sdk]    # Anthropic Claude Agent SDK

# Vector databases
pip install aigie[pinecone]            # Pinecone
pip install aigie[qdrant]              # Qdrant
pip install aigie[chromadb]            # ChromaDB
pip install aigie[weaviate]            # Weaviate
pip install aigie[vectordbs]           # All vector DBs

# Observability
pip install aigie[opentelemetry]       # OpenTelemetry inbound
pip install aigie[otlp]                # OTLP export

# Everything
pip install aigie[all]

Quick start

Decorator-based tracing (recommended)

from aigie import traceable

@traceable(run_type="agent")
async def my_agent(query: str):
    result = await process_query(query)
    return result

result = await my_agent("What is the weather?")

Auto-instrument LLM providers

from aigie import wrap_openai, wrap_anthropic
from openai import AsyncOpenAI
from anthropic import AsyncAnthropic

# OpenAI -- all calls automatically traced with model, tokens, cost, latency
client = wrap_openai(AsyncOpenAI())
response = await client.chat.completions.create(
    model="gpt-4",
    messages=[{"role": "user", "content": "Hello!"}],
)

# Anthropic
client = wrap_anthropic(AsyncAnthropic())

Context manager

from aigie import Aigie, Config

config = Config(
    api_url="https://api.aigie.com",
    api_key="your-key",
    batch_size=100,
    flush_interval=5.0,
)

aigie = Aigie(config=config)
await aigie.initialize()

async with aigie.trace("My Workflow") as trace:
    async with trace.span("operation", type="llm") as span:
        result = await do_work()
        span.set_output({"result": result})

Integrations

Agent frameworks

Framework	Install extra	Auto-instrument
LangChain	langchain	AigieCallbackHandler
LangGraph	langgraph	wrap_langgraph()
Claude Agent SDK	claude-agent-sdk	patch_claude_agent_sdk()

LLM providers

Provider	Wrapper
OpenAI	wrap_openai()
Anthropic	wrap_anthropic()
Google Gemini	wrap_gemini()
AWS Bedrock	wrap_bedrock()

Each integration lives in sdk/aigie/integrations/<framework>/ and follows a consistent pattern with auto-instrumentation, cost tracking, drift detection, error detection, retry logic, and session management. See CONTRIBUTING.md for details on adding new integrations.

Configuration

Environment variables

export AIGIE_API_URL=https://your-instance.aigie.io/api
export AIGIE_API_KEY=your-api-key
export AIGIE_BATCH_SIZE=100
export AIGIE_FLUSH_INTERVAL=5.0

Config object

from aigie import Config

config = Config(
    api_url="https://api.aigie.com",   # Aigie API endpoint
    api_key="your-key",                # API key
    batch_size=100,                    # Events per batch (default: 10)
    flush_interval=5.0,               # Flush interval in seconds
    enable_buffering=True,            # Enable event buffering (default: True)
    max_retries=3,                    # Retry count on failure
)

Advanced features

OpenTelemetry integration

from aigie import Aigie
from aigie.opentelemetry import setup_opentelemetry

aigie = Aigie()
await aigie.initialize()
setup_opentelemetry(aigie, service_name="my-service")

# All OTel spans now flow to Aigie
from opentelemetry import trace
tracer = trace.get_tracer(__name__)
with tracer.start_as_current_span("operation"):
    pass

Evaluation and scoring

from aigie import score, feedback

await score(trace_id, "accuracy", 0.95)
await feedback(trace_id, "user_feedback", "Great response!")

Prompt management

from aigie import Prompt

prompt = Prompt.chat(
    name="customer_support",
    messages=[{"role": "system", "content": "You are a helpful assistant."}],
    version="1.0",
)

Synchronous API

from aigie import AigieSync

aigie = AigieSync()
aigie.initialize()

with aigie.trace("workflow") as trace:
    with trace.span("operation") as span:
        result = do_work()
        span.set_output({"result": result})

Zero data retention

Wrap a call in aigie.no_retention() to keep its execution data inside your environment. Nothing about that run leaves it.

from contextlib import nullcontext

ctx = aigie.no_retention() if request.zero_retention else nullcontext()
with ctx:
    result = app.invoke(state, config)

Use aigie.no_retention_async() in async code.

API reference

Full API documentation is available at docs.aigie.io/sdk/python.

Development

# Clone and set up
git clone https://github.com/Kytte-AI/kytte-python-sdk.git
cd kytte-python-sdk
python -m venv .venv
source .venv/bin/activate
pip install -e "sdk/[dev]"

Common commands

make lint        # Run ruff linter
make format      # Format code with ruff
make test        # Run unit tests
make test-all    # Run all tests including integration
make coverage    # Run tests with coverage report
make typecheck   # Run mypy type checking
make check       # Run all checks (lint + test)
make build       # Build distribution packages

Running tests

# Unit tests
pytest tests/unit/ -v

# Integration tests (requires API keys)
pytest tests/integration/ -v

# Coverage report
pytest tests/unit/ --cov=sdk/aigie --cov-report=html --cov-report=term-missing

Publishing

Releases are published to PyPI automatically when a GitHub release is created, via the publish.yaml workflow.

Manual publishing is also supported:

./scripts/publish-sdk.sh <version>
# e.g. ./scripts/publish-sdk.sh 0.2.39

Contributing

See CONTRIBUTING.md for development setup and guidelines.

License

Proprietary -- Copyright (c) 2026 Kytte AI, Inc. All Rights Reserved. See LICENSE for details.

Documentation

• API Reference
• Integration Guides
• Changelog

Related

Repository	Description
docs-site	Documentation site

Website: https://kytte.ai/