Agent observability SDK for tracking AI agent execution
Project description
Cascade SDK
Agent observability platform for tracking AI agent execution, LLM calls, and tool usage.
Quick Start
Install from PyPI
pip install cascade-sdk
Setup
-
Set your API key (get one from Cascade Dashboard):
export CASCADE_API_KEY="your-api-key"
-
Use in your code:
from cascade import init_tracing, trace_run, wrap_llm_client, tool from anthropic import Anthropic import os # Initialize tracing (uses cloud endpoint by default) init_tracing(project="my_project") # Wrap LLM client client = wrap_llm_client(Anthropic(api_key=os.getenv("ANTHROPIC_API_KEY"))) # Decorate tools @tool def my_tool(query: str, client) -> str: """My custom tool.""" response = client.messages.create( model="claude-3-haiku-20240307", max_tokens=100, messages=[{"role": "user", "content": query}] ) return response.content[0].text # Trace agent execution with trace_run("MyAgent", metadata={"task": "example"}): result = my_tool("What is AI?", client) print(result)
-
View traces in the Cascade Dashboard
Features
- ✅ Zero setup - No backend services to run
- ✅ Cloud-first - Traces automatically sent to cloud
- ✅ LLM tracking - Automatic tracking of LLM calls (Anthropic, OpenAI, etc.)
- ✅ Tool tracing - Decorate functions with
@toolfor automatic tracing - ✅ Rich metadata - Add custom metadata to traces
- ✅ OpenTelemetry - Built on OpenTelemetry standards
Configuration
Environment Variables
CASCADE_API_KEY- Your Cascade API key (required)CASCADE_ENDPOINT- Override default endpoint (default:https://api.runcascade.com/v1/traces)
Custom Endpoint
If you need to use a custom endpoint:
init_tracing(
project="my_project",
endpoint="https://your-custom-endpoint.com/v1/traces",
api_key="your-api-key"
)
CLI Commands
cascade info- Show setup instructions and informationcascade --help- Show help messagecascade --version- Show version
Development
This is a monorepo containing:
cascade/- SDK package (published to PyPI)backend/- Backend service (not included in package)dashboard/- Frontend dashboard (not included in package)
Install SDK in Development Mode
pip install -e .
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Filter files by name, interpreter, ABI, and platform.
If you're not sure about the file name format, learn more about wheel file names.
Copy a direct link to the current filters
File details
Details for the file cascade_sdk-0.2.0b11.tar.gz.
File metadata
- Download URL: cascade_sdk-0.2.0b11.tar.gz
- Upload date:
- Size: 47.0 kB
- Tags: Source
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/6.2.0 CPython/3.11.13
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
a394da505775c7d5ca00d8bd67bc478be53a0198c53064a0837b350181c65d2a
|
|
| MD5 |
2c373ffe5771c6a58334436eaabad2b0
|
|
| BLAKE2b-256 |
50d05466214b5057f73593fd5786b4e293de56b85a89aa33c04589483199e481
|
File details
Details for the file cascade_sdk-0.2.0b11-py3-none-any.whl.
File metadata
- Download URL: cascade_sdk-0.2.0b11-py3-none-any.whl
- Upload date:
- Size: 54.7 kB
- Tags: Python 3
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/6.2.0 CPython/3.11.13
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
5437703dc042f1fdc9dc997f0fce7f1893b7a8bf2de644e52269853ee13ce9c7
|
|
| MD5 |
762dad858f9ea096d620511599c37643
|
|
| BLAKE2b-256 |
73c675e16be0f2cbb2df6e6bd3f8dc1f10f72dbf054d320cc724b5714ae08a01
|