SaaSActive

Weave

Evaluate single-turn and multi-turn AI agents in Weave using the agents workflow and EvaluationLogger, scored with an LLM judge.

Open source page ↗

Field note

What it does

Weave is a platform for tracing and evaluating AI agents. This guide shows how to trace an agent as a conversation, score runs with an LLM judge, compare agent versions, and score multi-turn conversations.

Capabilities

Available capabilities

Model EvaluationFeedback CollectionEvaluationVersion TrackingDataset ManagementOpenTelemetry CompatibilityScore with LLM judgeScoring FunctionsLangChain IntegrationGrow score into scorecardTrace agents as conversationsPython and TypeScript SupportLoggingParallelismLLM JudgesCachingWeave UIRetriesCompare agent versionsScore multi-turn conversationsTracingMetrics Collection

Tags

Tags

TracingLLMEvaluationAgent

Ways to use it

Ways to use it

sdk

python / weave

sdk

python / anthropic

sdk

Python / weave

sdk

TypeScript / weave

sdk

Python / https://docs.wandb.ai/weave

sdk

Python / weave

sdk

TypeScript / weave

Product features

Product features

​ API Overview

​ class Agent

​ class AgentState

​ class AnnotationSpec

​ class Audio

​ classmethod preprocess_field_schema

​ classmethod validate_field_schema

​ method export

​ method __init__

​ method step

​ method value_is_valid

weave