SDKActive

Weave SDK

Python and TypeScript libraries for tracing, evaluating, and improving LLM and agent applications.

Open source page ↗

Field note

What it does

Weave SDK provides Python and TypeScript libraries to instrument LLM calls and functions, trace agents, and collect feedback.

Capabilities

Available capabilities

Serverless deploymentFeedback CollectionEvaluationManage UsersMonitor AI applicationsDashboardControlServerless SFTSecurityObservabilityAgentsManage SecretsServerless InferenceLLM TracingEvaluate AI applicationsChatAI Coding AgentsSandboxesAI ModelsPrompt ManagementTrace AI applicationsDocumentationServerless RLInferenceAI AgentsServerless TrainingBuildTracingBilling

Tags

Tags

ObservabilityLLM observabilityAILLMEvaluation

Ways to use it

Ways to use it

pypi

python / weave

pypi

Python / weave

npm

TypeScript / weave

sdk

python / openai

rest

Not filed / https://docs.wandb.ai/weave

rest

Not filed / Not filed

Product features

Product features

Cost Tracking

Track costs for production AI apps.

Evaluation

Experiment tracking

Observability

Ops and Calls

Manually instrument LLM calls and arbitrary functions to trace, version, and collect feedback.

OTel-compatible SDK

Trace and collect metrics about agents built with popular SDKs and harnesses.

Python SDK

Python client for Weave tracing, evaluations, and trace server integration.

Scorers

Evaluate responses using LLM judges and custom scorers.

TypeScript SDK

TypeScript client for Weave, covering classes, functions, interfaces, and type aliases.

Weave Leaderboard

Evaluate final results on a Weave leaderboard.

Weave Playground

Explore and try supported models in the W&B Weave Playground.

Weave Scorers

Score and evaluate trained models during and after training.