Issues

PUT /api/scenarios/{id} returns 200 and silently drops unknown fields (status)
langwatch/tasks#925 · opened on Oct 3, 2026
1
POST /api/suites returns 201 and silently drops unknown fields (schedule, cron)
langwatch/tasks#924 · opened on Oct 3, 2026
1
An identical scenario write creates a new, empty version
langwatch/tasks#923 · opened on Oct 3, 2026
1
Grouped graph alerts add per-group averages together
langwatch/tasks#922 · opened on Oct 3, 2026
1
Monitor parameters and the saved evaluator's settings can disagree; only one is used
langwatch/tasks#921 · opened on Oct 3, 2026
2
PATCH /api/triggers/{id} returns 200 but ignores customGraphId
langwatch/tasks#920 · opened on Oct 3, 2026
1
Analytics metadata.key / metadata.value filters match nothing for SDK metadata
langwatch/tasks#919 · opened on Oct 3, 2026
2
Trace filter evaluator:X AND evaluatorVerdict:fail matches verdicts from other evaluators
langwatch/tasks#918 · opened on Oct 3, 2026
2
Tracking: platform gaps that keep experiment, scenario and alerting workflows off LangWatch
langwatch/tasks#917 · opened on Oct 2, 2026
1
Voice scenarios: slow turn-taking between simulated user and voice agent — map the turn, find the constraints
langwatch/tasks#916 · opened on Oct 2, 2026
Instant Evals as a judge model in evaluators and simulations, billed through the gateway
langwatch/tasks#915 · opened on Oct 1, 2026
Traces page search bar does not apply its query
Bug
langwatch/tasks#914 · opened on Sep 30, 2026
Keyed evaluation filters on trace search return 500
Bug
langwatch/tasks#913 · opened on Sep 30, 2026
Dashboards v2: clearer empty board, correct numbers, data freshness, and widgets from your own AI agent
langwatch/tasks#911 · opened on Sep 28, 2026
MCP /sse returns 405 on POST, breaking Claude Desktop; docs also point clients at /sse
langwatch/tasks#909 · opened on Sep 28, 2026
Tracking: API, MCP and trace-processing gaps (triggers, feedback filter, trace I/O, datasets)
Task
langwatch/tasks#908 · opened on Sep 28, 2026
Dashboards v1: Agent Flight Deck and question-first user dashboards on LangWatchQL
langwatch/tasks#907 · opened on Sep 28, 2026
EPIC: Dashboards on LangWatchQL, with the Agent Flight Deck as the curated default
langwatch/tasks#906 · opened on Sep 28, 2026
Benchmark datasets from real use: LLM-as-a-judge, agent simulation judge, and search
langwatch/tasks#905 · opened on Sep 25, 2026
LangWatchQL across projects in the product: cross-project dashboard charts and find a trace by id in any project
langwatch/tasks#904 · opened on Sep 24, 2026
Data privacy: count project and org role bindings in the content audience, and let API keys read restricted content they own
langwatch/tasks#903 · opened on Sep 24, 2026
Instant Evals as an option on LLM-as-a-judge evaluators (via our gateway, billed)
langwatch/tasks#902 · opened on Sep 24, 2026
Benchmark: best jailbreak and prompt-injection detectors, including Jev
langwatch/tasks#899 · opened on Sep 24, 2026
Benchmark: best PII detectors, multi-country dataset, including Jev
langwatch/tasks#898 · opened on Sep 24, 2026
Benchmark: Jev vs frontier LLM judges (gpt-5.6-luna, Gemini, Opus and others)
langwatch/tasks#897 · opened on Sep 24, 2026