Integrations¶

Add one line to your agent, and every run is recorded and diagnosed. The SDKs have no dependencies.

pip install agent-tracedown          # Python
npm install agent-tracedown          # TypeScript

import agent_tracedown as tdn
tdn.init(project="support-bot", endpoint="http://127.0.0.1:7070")

Frameworks¶

Framework	One line
LangChain, LangGraph	`graph.invoke(x, config={"callbacks": [tdn_lc.callback("MyGraph")]})`
OpenAI Agents SDK	`tdn_oa.instrument()`
Claude Agent SDK	`async for m in tdn_claude.trace_stream(query(prompt=p, options=o), name="Agent")`
Google ADK	`App(name="support", root_agent=agent, plugins=[tdn_adk.TracedownPlugin()])`
Pydantic AI	`result = await tdn_pai.run(agent, prompt)`
MCP clients	`tdn_mcp.instrument(client)`
CrewAI	`tdn_crew.instrument()`
AutoGen AgentChat	`async for m in tdn_ag.trace_stream(team.run_stream(task=t), name="Team")`
Semantic Kernel	`tdn_sk.instrument(kernel)`
LlamaIndex	`tdn_li.instrument()`
OpenAI and Anthropic clients	`tdn.plugins.instrument("openai", client)`
Vercel AI SDK (TypeScript)	`generateText({ model: traceModel(model), tools: traceTools(tools) })`
OpenAI Agents SDK (TypeScript)	`await instrument()` from `agent-tracedown/openai-agents`

Python adapters live in agent_tracedown.integrations. Each one is tested in CI against the real library.

Without a framework¶

with tdn.run("CalendarAssistant", input=message) as run:
    reply = tdn.llm_call(call_model, model="gpt-4o-mini", prompt=messages)
    tdn.tool_call("create_event", args, create_event, schema=SCHEMA)
    run.output(reply)

The SDK READMEs cover every call, redaction and replay: Python · TypeScript.

Exports¶

Tracedown also sends spans on as OTLP (Jaeger, Tempo, Datadog or any collector) and serves Prometheus metrics at /metrics. See Configuration.