blockwatch 0.1.2

Lightweight telemetry for Amazon Bedrock multi-agent collaboration and orchestration

BlockWatch Python SDK

Non-blocking telemetry for agent orchestration (Bedrock today; LangGraph, CrewAI, AutoGen planned). Traces are collected in the background and sent to BlockWatch — your app code stays unchanged.

Install: pip install blockwatch · PyPI: https://pypi.org/project/blockwatch/

Quick start

1. Install the package.
2. Call init() once with your API key and framework.
3. For Bedrock, use bedrock-agent-runtime as usual (before creating the client).

pip install blockwatch

import os
import blockwatch
import boto3

blockwatch.init(
    api_key=os.environ["BLOCKWATCH_API_KEY"],
    framework="bedrock",
)

client = boto3.client("bedrock-agent-runtime", region_name="us-east-1")

response = client.invoke_agent(
    agentId="YOUR_AGENT_ID",
    agentAliasId="YOUR_ALIAS_ID",
    sessionId="your-session-id",
    inputText="Hello",
)

for event in response.get("completion", []):
    if "chunk" in event:
        print(event["chunk"]["bytes"].decode(), end="")

enableTrace=True is set for you automatically when framework="bedrock".

AWS Lambda

requirements.txt

blockwatch>=0.1.2

Call blockwatch.init(api_key=..., framework="bedrock") at module load. Build your layer on Amazon Linux:

pip install -r requirements.txt -t python/

API

Function	Description
blockwatch.init(api_key, framework)	Start telemetry. framework: bedrock, langgraph, crewai, or autogen.

Telemetry is sent to https://api.blockwatch.ai/telemetry with header X-BlockWatch-Framework set to your framework value.

Notes

• Python 3.9+
• Requires boto3 and requests (installed with the package)
• SDK errors never crash your app (failures are silent)
• Boto3 patching applies only when framework="bedrock"