Skip to content

Prodgy Overview

Latest Version v0.0.13 Stable

Section titled “Latest Version ”

Native Prodgy Dashboards for organizations powered by internal events

New Organizations now get a set of native dashboards installed automatically — MAU, DAU, Stickiness, AI consumption, top users, active workspaces and more — fed by internal events and LLM usage instead of external integrations. Admins and Superadmins see business adoption and cost without leaving Prodgy.

Superadmin platform health and cross-org analytics

New A dedicated set of cross-organization dashboards for the Prodgy internal team: platform health, feature adoption, LLM cost by source/model/integration, top MCP tools, top chat users, meetings frequency, user feedback aggregates and an individual user activity drill-down.

Per-workspace dashboards for product analytics

New Every workspace receives its own dashboards: distinct users, installed agents, embedded chunks, knowledge base size, agent executions and errors, active integrations, top agents over time, playground usage, top users and features-per-user breakdown.

Agent-owned dashboards bundled with the agent

New Agent templates can now declare their own dashboards that are installed alongside the agent into the target workspace and removed on uninstall — the same pattern integrations already use. Meeting Scriber ships as the pilot, with Scheduled Meetings, Recorded Meetings, Completed Executions and Completed Meetings dashboards.

Dashboard templates catalog and definition management

New A central catalog separates dashboard templates from organization/workspace instances, classified by origin (integration, agent, platform) and scope (superadmin, organization, product). New templates are materialized automatically on organization/workspace creation and on agent install.

Customizable dashboard layouts, tabs and user personalization

New Users can organize dashboards into tabs, drag and drop into resizable slots, save the current layout as default for their organization or workspace, and reset preferences at any time. A three-layer cascade (user → org/workspace → platform default) keeps consistency across teams.

Customizable dashboard layouts with slot-based drag-and-drop

New Dashboard slots support multiple sizes (small, medium, large, full), drag-and-drop reordering and form-factor compatibility (tile, graph, table). Suggested layouts and free-grid editing let each user shape the canvas to their workflow.

Hierarchical scope navigation for superadmin, admin and user roles

New The navigation now reflects the user role: superadmin sees the platform view, admins land on the organization view, and only after selecting a workspace does the product context appear — with a clear path back to the broader scope.

Migrate Prodgy Assistant portal into the Prodgy admin area

New The Prodgy Assistant operational dashboard (Schedule, Logs, History) is now embedded in the main Prodgy admin area, removing the need to access a separate web portal.

Improved retrieval of meetings and documents in the Prodgy chat

New Chat-based retrieval (Communicator and Playground) now uses structured filters, facets and groundedness controls to bring more accurate answers grounded in meetings and documents.

Meeting Summarization email template for the Meeting Scriber agent

New Meetings transcribed by the Meeting Scriber agent now generate a polished summary email automatically, ready to be sent to the organizer and attendees.

AI Models admin tab and default provider/model display in Playground

New A new AI Models tab in the admin area lets superadmins see and manage available LLM providers and models. The Playground now shows the currently selected default provider/model up front.

Join meetings via @Prodgy mention directly in Teams chat

New In addition to scheduled email invites, users can now bring the Prodgy Assistant into an ongoing Teams meeting just by mentioning @Prodgy join in the meeting chat — the bot validates the requester, resolves the right workspace and starts transcribing in real time.

Better @Prodgy chat command UX with auto-workspace and privacy notice

New When triggered via chat command, the bot remembers each user’s pinned workspace, avoids friction when multiple workspaces are accessible, and posts a clear privacy/LGPD notice when entering the meeting.

Search users in the user list

New The user management list now has a search box, so admins can find a user by name or email without scrolling the entire directory.

Manage users in workspace settings (same UX as the add-workspace flow)

New The “Users” tab inside workspace settings now matches the rich UX of the add-workspace screen — including search, role selection and batch additions.

Self-service Microsoft Teams installation in Prodgy

New Customers can now install and configure the Microsoft Teams integration through a guided self-service flow inside Prodgy, with no manual admin steps required from the Prodgy team.

Communicator bot repositioned as Prodgy AI Brain

New The Communicator bot now appears across surfaces as “Prodgy AI Brain”, with refreshed naming and presentation to make it clearer that it’s the AI assistant tied to the workspace’s knowledge and agents.

Centralized public documentation and legal URLs via backend configuration

New Privacy, Terms and Docs URLs are now served by the backend as a single source of truth — any future change rolls out instantly across the app, the Communicator and the Assistant without redeploys.

Inline actions on chat history (delete/archive/rename) with full history access

New Past conversations in the Playground can now be renamed, archived or deleted inline. The chat history list shows the full history, not just the latest sessions.

Core MCP tool get_my_profile executed by default before project context

New AI agents acting on the user’s behalf automatically know “who am I” at the start of each conversation, with name, role and workspace pinned in context — producing better-personalized responses.

Knowledge base hybrid search enriched with dynamic JSONB metadata

New Hybrid search across the knowledge base now weighs document metadata fields dynamically, improving precision on domain-specific terms, names and identifiers without manual tuning.

Faster chat responses via core LLM query classification

New A lightweight classifier in front of the chat decides whether a query needs retrieval, tools or a direct answer — cutting unnecessary round-trips and producing noticeably faster responses on simple questions.

Playground AI honors the same MCP server contract as external clients

New The Playground assistant now talks to integrations and agents through the same MCP contract used by external clients (Claude.ai, Cursor, ChatGPT), keeping behavior consistent and reducing duplicated wiring.

Separated AI thinking trace from the final response in Playground and Communicator

New Reasoning and tool-use traces are now clearly separated from the final assistant response in both the Playground and Communicator, making conversations easier to read while keeping debugging context available on demand.