AgentTrace SDK
AI Agent Observability & Debugging - Record, replay, and debug your AI agents.
🚀 Quick Start
pip install agenttrace-sdk
from agenttrace import AgentTrace
# Initialize with your API key
tracer = AgentTrace(api_key="at_your_api_key_here")
# Start a trace session
tracer.start("my_agent.py")
# Record agent activities
tracer.thought("Analyzing the user's request...")
tracer.tool_start("web_search", {"query": "latest AI news"})
tracer.tool_end("web_search", {"results": [...]})
# End the session
tracer.end()
📋 Features
- 📍 Event Recording - Capture thoughts, tool calls, file writes, and exceptions
- 🔄 Replay Execution - Re-run your agent with deterministic results
- 🌿 Branch & Fork - Create alternate timelines from any point
- 🔍 Compare Runs - Side-by-side diff of different executions
- 🛠️ AutoFix Engine - AI-powered error detection and fixes
🔧 API Reference
Initialize Client
from agenttrace import AgentTrace
tracer = AgentTrace(
api_key="at_xxx...", # Your API key (required)
endpoint="https://app.agenttrace.io", # API endpoint (optional)
batch_size=10, # Events before auto-flush
flush_interval=1.0 # Seconds between flushes
)
Record Events
# Record a thought/reasoning step
tracer.thought("Planning the next action...")
# Record tool invocation
tracer.tool_start("calculator", {"expression": "2+2"})
tracer.tool_end("calculator", result=4)
# Record file operations
tracer.file_write("/path/to/file.txt", "content here")
# Record exceptions
tracer.exception("ValueError", "Invalid input", traceback_str)
Context Manager
with tracer.session("my_script.py") as trace_id:
tracer.thought("Starting...")
# Your agent code here
# Automatically ends with status "completed" or "failed"
Decorator
from agenttrace import trace
@trace
def my_function(x, y):
return x + y # Automatically traced!
🔑 Get Your API Key
- Sign up at agenttrace.io
- Go to Settings → API Keys
- Create a new key
- Copy and use in your code
# Or set via environment variable
export AGENTTRACE_API_KEY="at_your_key_here"
# Then just:
tracer = AgentTrace() # Auto-reads from env
📊 View Your Traces
After recording, view your traces at:
- Dashboard:
https://app.agenttrace.io/dashboard/traces - Individual Trace:
https://app.agenttrace.io/trace/{trace_id}
🌐 Self-Hosted
For self-hosted deployments:
tracer = AgentTrace(
api_key="at_xxx...",
endpoint="https://your-domain.com" # Your self-hosted instance
)
📄 License
MIT License - see LICENSE for details.
🔗 Links
- Website: agenttrace.io
- Documentation: docs.agenttrace.io
- Support: agenttraceismoat@gmail.com
Metadata
Release files for agenttrace-sdk 1.0.0
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | Provenance | Actions |
|---|---|---|---|---|
| agenttrace_sdk-1.0.0.tar.gz | 9.0 kB | Without provenance | Details for agenttrace_sdk-1.0.0.tar.gz |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Actions Reset |
|---|---|---|---|---|
| agenttrace_sdk-1.0.0-py3-none-any.whl | Python 3 | none | any | Details for agenttrace_sdk-1.0.0-py3-none-any.whl |
Total release size: 16.8 kB
Release files / agenttrace_sdk-1.0.0.tar.gz
| Download URL | agenttrace_sdk-1.0.0.tar.gz |
|---|---|
| Size | 9.0 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
c6e7f1163d89e77a9850886b5dd8d69eaf115d214d082afb965f4b848bcbd5d9
|
|
BLAKE2b-256 checksum How to use checksums |
bef01ce7b47398dd11ee6068ed54d6d8e7a4cfb0ea72c3e8b32470c155f7f58b
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/6.2.0 CPython/3.12.1
|
Release files / agenttrace_sdk-1.0.0-py3-none-any.whl
| Download URL | agenttrace_sdk-1.0.0-py3-none-any.whl |
|---|---|
| Size | 7.8 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
b0f9ec63bc00d29fbe6d0f88a279d80844ac139452d5f814b2b7216bf982e152
|
|
BLAKE2b-256 checksum How to use checksums |
7bfd57b0ee419947f186b18dc853dc138ad8dcf29be59f1f1b73c34948b57843
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/6.2.0 CPython/3.12.1
|
Security for agenttrace-sdk 1.0.0
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
No provenance found
None of the files in this release were uploaded with provenance metadata.
A note on package security
PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.
To learn more about protecting yourself from malicious actors, visit our package security documentation.