Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.langsight.dev/llms.txt

Use this file to discover all available pages before exploring further.

Wrap an Anthropic or AsyncAnthropic client with wrap_llm() to auto-trace every messages.create() call and tool_use block in the response.
Using Claude Agent SDK? See the dedicated Anthropic integration — LangSightClaudeAgentHooks covers the full agent runner lifecycle.

​
Installation

pip install langsight anthropic

​
Quick start

import langsight
from anthropic import Anthropic

ls = langsight.init()

raw_client = Anthropic()
client = ls.wrap_llm(raw_client, agent_name="my-agent", session_id="sess-001")

response = client.messages.create(
    model="claude-sonnet-4-6",
    max_tokens=1024,
    tools=[{"name": "get_weather", "description": "...", "input_schema": {...}}],
    messages=[{"role": "user", "content": "What's the weather in Berlin?"}],
)
# response is unchanged — same Anthropic response object
# tool_use content blocks auto-traced as tool_call spans
The wrapper intercepts client.messages.create() (sync) and client.messages.acreate() (async).

​
What gets traced

SpanFields
LLM generationserver_name="anthropic", tool_name="generate/claude-sonnet-4-6", span_type="agent", tokens, model_id
Tool use (per block)tool_name from response, span_type="tool_call", parent_span_id, input_args
Session: sess-001
├── anthropic/generate/claude-sonnet-4-6  800ms  success  (200 in / 80 out tokens)
│   └── get_weather                                       success

​
Parameters

client = ls.wrap_llm(
    Anthropic(),               # required — anthropic.Anthropic or anthropic.AsyncAnthropic
    agent_name="my-agent",     # optional — shown in dashboard
    session_id="sess-001",     # optional — groups spans into a session
    trace_id="trace-abc",      # optional — links across multi-agent tasks
)