Lunary vs W&B Weave
Compare Lunary and W&B Weave for agent conversations, feedback APIs, prompt versions and evaluation. See which fits a customer-facing AI product.
$20 per user / month. Payment required to activate your workspace.
Lunary is our pick for application teams that want to connect what a user saw, how they reacted, and what the agent actually did.
Lunary fit
Conversation replay with application feedback.
Weave fit
Agent tracing and versioned evaluation artifacts in W&B.
Check SDK parity
Confirm feedback and object operations in your language.
At a glance
| Workflow | Lunary | W&B Weave |
|---|---|---|
| Review interactions | Chat replay with linked model and agent runs | Agent sessions and turns |
| Capture feedback in code | JavaScript and Python APIs | Python API; TypeScript unsupported in current guide |
| Version prompts | Templates with drafts and team review | Versioned objects, tags and aliases |
| Compare quality | Evaluate models and prompt configurations | Dataset evaluations and scoring functions |
See your workflow in Lunary.
Explore conversations, traces and prompt versions with our team.
Review the customer experience
In Lunary, messages can link to LLM calls and agent runs, with custom events and feedback alongside the interaction. Weave also has an Agents view for sessions and turns, plus a separate Ops and Calls tracing workflow.
Check the code you need to ship
Lunary provides JavaScript and Python feedback examples for frontend or backend capture. Weave’s feedback guide documents its UI and Python SDK, with TypeScript support marked unavailable. That distinction matters if feedback is part of your TypeScript application.
Make prompt changes reviewable
Lunary combines versioned templates with draft sharing and a playground. Weave versions prompts, datasets and model configurations as objects. Trial the edit, review and release process your team will repeat each week.
Questions & answers
Is Weave only for ML experiments?
No. Weave provides agent observability, LLM tracing, evaluations and feedback.
Which should a TypeScript team try first?
We recommend Lunary if collecting in-app feedback is essential. Validate the exact SDK operations your application needs on both platforms.
What would a useful Lunary pilot prove?
That your team can locate a poor reply, inspect its trace, review a prompt revision and assess the change on representative examples.
By Lunary · Documentation reviewed Sep 19, 2026
Sources & methodology (9)
Based on vendor documentation, not an independent benchmark or hands-on rating. Features and plans can change.
Bring your next AI issue to Lunary.
Follow the conversation. Find the cause. Test the fix.
$20 per user / month. Payment required to activate your workspace.