# wandb/weave

Weave is a toolkit for developing AI-powered applications, built by Weights & Biases.

Repository: https://github.com/wandb/weave
Canonical: https://ross.abutalabs.com/products/wandb-weave
Homepage: https://wandb.me/weave
Language: Python
License: Apache-2.0
License Family: permissive
Last push: 2026-09-02T21:42:02+00:00

## Health v2 (maintenance only)
Score: 92/100 (v2, computed 2026-09-03T02:39:23.370411+00:00)
- activity 100, release rhythm 87, longevity 84
- inputs: {"age_days": 1184, "days_push": 0, "days_rel": 6, "gap_med": 6, "n_releases_24m": 88}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1124, forks 166 (observed 2026-09-03T02:15:20.431163+00:00)

## What it is
Weave is a Python/TypeScript toolkit by Weights & Biases for developing AI-powered applications, providing LLM tracing, agent observability, and evaluation tooling. It instruments LLM calls and arbitrary functions via decorators and OTel-compatible spans, logging inputs, outputs, and traces to the W&B platform.

## Use cases
- trace and debug LLM calls and function inputs/outputs
- observe multi-turn agents built with OpenAI Agents SDK or Claude Agent SDK
- evaluate LLM app responses with LLM judges and custom scorers
- evaluate a RAG application
- send OpenTelemetry spans from an agent to a UI
- collect metrics and feedback about LLM applications in production

## When to choose
- you need observability and tracing for LLM apps or agents
- you want rigorous, comparable evaluations with scorers and LLM judges
- you already use the Weights & Biases ecosystem
- you want autopatching integrations for popular agent SDKs

## When to avoid
- you need fully self-hosted telemetry with no external service dependency
- your project is not LLM/AI related
- you cannot send data to the W&B cloud platform

## Facets
- artifact type: library
- maturity: active
- function: monitoring, tracing, testing, analytics, llm-inference, agent-framework
- domain: large-language-models, machine-learning, developer-tools, monitoring
- platform: python, cross-platform
- tags: llm-observability, llm-evaluation, tracing, opentelemetry, weights-and-biases, genai-toolkit, ai-agents, nodejs

## Member repositories
- wandb/weave (main) score 92

## Provenance
- Observed fields: from GitHub, fetched 2026-09-03T02:15:20.431163+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T06:41:02.761535+00:00, confidence not recorded.
  - readme: https://github.com/wandb/weave (fetched 2026-09-03T02:15:20.431163+00:00, sha 4c5e58ab8dfb)
  - homepage: https://wandb.me/weave (fetched 2026-08-29T12:45:07.796312+00:00, sha 334cb2ed3343)
  - site_page: https://docs.wandb.ai/ (fetched 2026-08-29T12:45:07.800173+00:00, sha 8c6fe7d8e95e)
  - site_page: https://docs.wandb.ai/weave/agent-integration-quickstart (fetched 2026-08-29T12:45:07.802472+00:00, sha d0ae6029671a)
  - site_page: https://docs.wandb.ai/weave/custom-agents-quickstart (fetched 2026-08-29T12:45:07.804177+00:00, sha 8dcf023f6a5d)
  - site_page: https://docs.wandb.ai/weave/quickstart (fetched 2026-08-29T12:45:07.806129+00:00, sha 745e130017f1)
  - registry_pypi: https://pypi.org/pypi/weave/json (fetched 2026-08-29T12:45:07.807812+00:00, sha c1f349781c02)
- Data as of 2026-08-30T08:39:29.467469+00:00.