agenta
agenta copied to clipboard
Agenta-AI
The open-source LLMOps platform: prompt playground, prompt management, LLM evaluation, and LLM observability all in one place.
@mmabrouk said: > The container for load application in evaluator playground has the wrong height [CleanShot 2025-11-05 at 15.04.41.mp4](https://uploads.linear.app/514b189b-9687-46df-a5d7-242a42e958ea/3b6f1a12-d118-4545-90d6-4590eaf44915/5ccf3f5e-3f4b-4608-9f77-fbdcf2ff04f8) From [SyncLinear.com](https://synclinear.com) | [AGE-3376](https://linear.app/agenta/issue/AGE-3376/fix-incorrect-height-of-load-application-container-in-evaluator)
> It happens when you go from evaluations to the playground. > My guess is that it's not the only page change that breaks the breadcrumbs From [SyncLinear.com](https://synclinear.com) | [AGE-3375](https://linear.app/agenta/issue/AGE-3375/bug-the-breadcrumb-is-sometimes-wrong)
**Labels**: bug, authentication, self-hosted, invitation --- ### **Bug Summary** Invited users can successfully create accounts but receive 401 Unauthorized errors when attempting to access workspace resources, preventing them from joining...
@mmabrouk said: > \[Minor\]\[UI\] Overlap text in the Playground in chat mode 🧵 [CleanShot 2025-10-30 at 11.04.59.mp4](https://uploads.linear.app/514b189b-9687-46df-a5d7-242a42e958ea/e7b9ea92-7b03-491f-97ad-c7063684da57/cf34ce30-d038-4f10-8369-62307500c45a) From [SyncLinear.com](https://synclinear.com) | [AGE-3362](https://linear.app/agenta/issue/AGE-3362/bug-text-overlap-issue-in-playground-chat-mode-ui)
**Describe the bug** There is an issue in the evaluator Playground in the editor which makes it so that if you provide a string that has a JSON Markdown, it...
### Discussed in https://github.com/Agenta-AI/agenta/discussions/2699 Originally posted by **mmabrouk** July 24, 2025 The traces shown in observability include references to the: - Application used - Variant used - Version used -...
@mmabrouk said: > In logfire adapter we are not correctly mapping tool definition (`model_request_parameters`) to their place. [example](https://cloud.agenta.ai/w/01909171-21d9-7c8c-8eef-f407067b1a5a/p/01927b4a-0d30-74cc-bd45-f73c3b19fadc/observability?trace=39e46ba7e412d65d7f555f0e64f1e4c6&span=dc566bc7960a2e7f) 🧵 From [SyncLinear.com](https://synclinear.com) | [AGE-3312](https://linear.app/agenta/issue/AGE-3312/in-logfire-adapter-we-are-not-correctly-mapping-tool-definition-model)
https://agenta-hq.slack.com/archives/C05JDQWKD6E/p1757609219363379 [AGE-3229](https://linear.app/agenta/issue/AGE-3229/bug-if-a-test-case-fail-during-eval-overall-statistics-are-not-shown)
For reference: https://ai.google.dev/gemini-api/docs/google-search#rest We can use google search simply by add ``` { "google_search": {} } ``` to tools. But in agenta it seems unable to load the tools due...
**Describe the bug** It is not clear which outputs are compared in the case of exact match evaluator when run on chat applications. The behavior and the best practices are...