Versions and releases

Citizen.chat has undergone multiple updates since its initial version. Here's a summary of the updates:

• v0.1: The initial version was released on April 15th.

• v0.2: This update included UI enhancements and new features, delivered on April 17th.

• v0.3: Addition of permissions and new features for roles/tasks, delivered on May 3rd. The update also introduced evaluations on new models (Mixtral8x22b, WizardLM2-8x22b).

• v0.4: This version was oriented toward assistant functionality, providing full support for more models (Mixtral8x22b, WizardLM2-8x22b, gpt-o, claude, gemini), while focusing on 14 roles and tasks design, delivered on May 6th.

Citizen.chat has outlined plans for several upcoming versions, each introducing new features and enhancements. Here is what's in the pipeline:

• v0.5: This version will include an upgrade to the UI based on user feedback and focus on resolving minor bugs. It will also introduce the feature of "my agents" accessible across datasets, improving workflow efficiency by reducing friction in use.

• v0.6: The v0.6 release will incorporate a multi-retriever feature, allowing the selection of data sources directly from the chat interface. A highlight will be the ability to attach one or more datasets to agents, enabling agents to query datasets and dynamically define guidelines based on the user's question.

• v0.7: This iteration will feature "teams of agents," wherein a planning agent automatically orchestrates different roles based on the question asked. This step towards automation is poised to enhance the platform's efficiency and effectiveness.

These updates are part of Citizen.chat's commitment to continuously improve and refine its platform in line with user feedback and the evolving landscape of AI functionalities.