Agentica
Agentica is an open-source AI coding assistant with installation links for VS Code, Cursor, and Windsurf. The publisher labels it early beta and warns of rough edges. Its main choices concern which model you use and whether access comes through a GenLabs plan or your own compatible endpoint.
Editor access and model choice
The site provides OpenVSX installation access alongside editor-specific links. It describes source code that users can inspect and modify, and separates free models from premium models in its catalog.
Agentica supports OpenAI-compatible endpoints. The publisher names Ollama and LM Studio among the options, as well as hosted providers or an endpoint you operate. This gives users with an existing model setup a route to connect it, although the homepage does not document every endpoint's behavior or compatibility limits.
Before choosing an endpoint, check that the model and access method you want appear in the current interface. The listed model catalog can change, and the free catalog and paid catalog have different access rules.
Daily requests and paid credits
The Free plan lists 100 free-model requests per day. The paid plans list unlimited open-model requests plus daily credits for premium-model usage. Plus is displayed at $15 per month with $1 in daily API credits, Pro at $50 with $4 daily, and Max at $200 with $10 daily.
Unused plan credits reset each day. A monthly subscription therefore does not mean you can accumulate the daily allocation for a large job later. Premium-model requests consume credits based on usage, so count both request limits and model costs when comparing access options.
Evaluate the beta on your coding workflow
The page requires a GenLabs account for its plans and offers cancellation from the dashboard. The captured homepage emphasizes model access and pricing but provides little detail on specific coding operations or repository limits.
Try the extension on a small, reviewable coding task before depending on it for a larger change. Check what context it receives, inspect suggested edits, and confirm how your selected model's usage affects the daily balance. Keep those observations separate from the publisher's comparative value slogans.