🪢 Open source agent evals & observability: Trace, evaluate, and improve LLM applications with one open platform.
TypeScript analyticsautogenevaluationlangchainlarge-language-modelsllama-indexllmllm-evaluationllm-observabilityllmopsmonitoringobservabilityopen-sourceopenaiplaygroundprompt-engineeringprompt-managementself-hostedycombinator
Commits on main
chore(agents): apply cursor label to Cursor PRs (#17165)
0e0458a Nikita Kabardin committed 2 months ago
fix(shared): align in-memory null filters with column semantics (#17160)
1bb49ba Tobias Wochinger committed 2 months ago
fix(web): Create `MultiSelectInput` & update `NewDatasetItemForm` (#17139)
4c63fad Ben Bachem committed 2 months ago
fix(evals): use radio controls for sample observation selection (#17134)
87ad135 Tobias Wochinger committed 2 months ago
fix(otel): normalize official cache-write usage (#17129)
ddf4434 Hassieb Pakzad committed 2 months ago
fix(seeder): write the evaluator columns in the bulk scores insert (#17131)
f3475e4 Lysander Kiesel committed 2 months ago
fix(ci): keep runtime report in issue output (#17132)
233f371 Tobias Wochinger committed 2 months ago
fix(auth): stop NextAuth email callback from consuming password-reset codes (#17039)
56154bd Max Deichmann committed 2 months ago
chore(ci): increase Blacksmith CPUs and tune concurrency (#17122)
aa5f60a Nimar committed 2 months ago
chore(pricing): document Sept 6 audit re-confirmation and new Gemini agent/media model wave (#17105)
c65c378 langfuse-bot committed 2 months ago
test(web): tolerate concurrent API key cache entries (#17123)
1694988 Tobias Wochinger committed 2 months ago
fix(playground): increase the output token limit ceiling to 65535 (#17038)
363b083 Thanos committed 2 months ago
chore(assistant): refresh embedded Langfuse skills (#17072)
26f15d3 langfuse-bot committed 2 months ago
feat(api)!: remove unstable evaluation endpoints (#17056)
619dbe3 Tobias Wochinger committed 2 months ago
refactor(experiments): say baseline and not-scored, and take a number off the header line (#17009)
7637df1 Nikita Kabardin committed 2 months ago
fix(experiments): make a long output readable in place, and say that it scrolls (#17007)
2de6a80 Nikita Kabardin committed 2 months ago
fix(web): keep the localStorage preference updater pure (#17006)
a646c24 Nikita Kabardin committed 2 months ago
feat(traces): show cost source in breakdown tooltip (#17051)
74c6064 Hassieb Pakzad committed 2 months ago
feat(analytics): instrument the rebuilt experiments UI (#16919)
9b850ab Nikita Kabardin committed 2 months ago
docs(api): advertise self-hosted API reference (#17084)
b1290b4 Tobias Wochinger committed 2 months ago
feat(experiments): a worse-than-comparison filter and a score-by-run matrix (#16918)
79b2c66 Nikita Kabardin committed 2 months ago
docs(agents): make verification proportionate to risk (#17064)
fc73049 Nikita Kabardin committed 2 months ago
docs(agents): rank by the handbook, and stop inventing staleness (#17075)
51e2bfe Nikita Kabardin committed 2 months ago
feat(experiments): rebuild the results grid around the score columns (#16917)
afde538 Nikita Kabardin committed 2 months ago