> ## Documentation Index
> Fetch the complete documentation index at: https://docs.clarityq.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Changelog

> Product updates and release notes for ClarityQ, the agentic data analytics platform. Newest first.

**November 12, 2026**

## Your Context Layer, now versioned with Git

Your Context Layer is now managed like code - every change is tracked, tested on its own branch, and reversible. Here's what that means for your team:

**Work in parallel, each on your own branch.** No more workspace-wide lock. Admins and Analyst-Contributors each get a private branch, can run several Builder chats at once, and use the Published / Your branch switch to see how the agent will answer before deploying.

**Conflicts resolved in the Builder.** When changes overlap, the Builder explains both versions and proposes a merged one to approve.

**Easy to undo.** Delete replaces Exclude, and anything deleted can be restored by asking the Builder. The Versions page lets you roll back any change.

### Changes in the ClarityQ Semantic Catalog: simpler events and parameters

To match the language your team already uses, features are now called events, and feature dimensions are now parameters. Updates to events and parameters, like rules or custom behavior, are now made in the Context Builder with the new /event and /parameter skills, instead of editing SQL or using the rule editor.

**November 8, 2026**

## Coverage badge and the Inbox, now available to everyone

### See how solid the context behind each answer is

Every ClarityQ answer now carries a context coverage badge, showing how complete the Context Layer was for that question - green when everything used is covered and approved (the vast majority, with a solid Context Layer), orange when unverified knowledge was used, red when some knowledge was missing. For every such gap, the agent suggests an update to the context - and you simply approve or edit the suggestion - so similar questions are answered with full confidence, the same way, every time.

### Meet the Inbox - your Context Layer to-do list

All detected gaps land in the new Inbox page, prioritized by what matters most: the type of gap, how many chats it appeared in, and whether users actively reported it. From there, Analysts and Admins can act in one click - approve or edit unverified components, add missing ones in the Builder, and review correction claims from the team. Your Context Layer keeps getting better, driven by real usage.

**October 18, 2026**

## Ask about anything in a dashboard, and ask ClarityQ from Claude, Cursor or ChatGPT

### Click anything in a dashboard or artifact and ask about it

You can now click any chart, number or table in an artifact or dashboard and ask ClarityQ about it right there. Ask why a number went up or what is driving a trend, or ask for a change, such as a different breakdown, chart type or layout, and ClarityQ answers or updates the dashboard or artifact in place.

### See how much to trust every description, and where it came from

Every description in the Event Catalog, Table Catalog and Semantic Catalog now shows its confidence and source right next to the text. Metrics, segments, skills and entities show their source (when known). ClarityQ also fills in descriptions for new events, parameters and user properties on its own and marks how sure it is, so you only need to review the ones marked Inferred. Approving or editing a description makes it Verified with your name on it. Open any catalog item to see it.

### Ask ClarityQ from Claude, Cursor or ChatGPT

You can now connect ClarityQ to MCP clients such as Claude, Cursor or ChatGPT and ask questions there, with the same answers and the same data permissions as in the app. Each member connects with their own ClarityQ account, and these chats appear in your chat history with an MCP icon. Ask your ClarityQ Admins to enable MCP for your organization first.

### Upload more kinds of files

Chat, the Builder and the Files page now accept the same file types, and the list has grown: documents (PDF, Word, PowerPoint, Excel), images, notebooks and source code in Python, JavaScript, TypeScript, Java, Go, Rust and more. If a file is not supported, the message tells you exactly which types were rejected.

**September 30, 2026**

## Coverage badge, the Inbox, and ClarityQ on mobile

### Start working with your Context Layer right away

No more waiting for every component to be verified before your team can ask questions. Your Context Layer components - including pending ones - are now available to the agent immediately, so you can start getting answers from day one and close gaps as they surface, instead of upfront.

### See how solid the context behind each answer is

Every ClarityQ answer now carries a context coverage badge, showing how complete the Context Layer was for that question - green when everything used is covered and approved (the vast majority, with a solid Context Layer), orange when unverified knowledge was used, red when some knowledge was missing. For every such gap, the agent suggests an update to the context - and you simply approve or edit the suggestion - so similar questions are answered with full confidence, the same way, every time.

### Meet the Inbox - your Context Layer to-do list

All detected gaps land in the new Inbox page, prioritized by what matters most: the type of gap, how many chats it appeared in, and whether users actively reported it. From there, Analysts and Admins can act in one click - approve or edit unverified components, add missing ones in the Builder, and review correction claims from the team. Your Context Layer keeps getting better, driven by real usage.

### ClarityQ, now on mobile

The chat experience is now supported on mobile - ask questions, follow up, and get answers wherever you are.

### See your prompt usage

A new Usage page gives you visibility into your prompts - track how much your team is asking and stay on top of your usage.

**August 30, 2026**

## Files are here, and introducing ClarityQ Tokens

### Files are here

You can now bring files into ClarityQ - and put them to work. Upload a CSV with additional data, an image of a dashboard you want to replicate, catalogs you maintain in files, or a presentation to use as design inspiration. Add files to your Context Layer to give the agent lasting knowledge across every session, or attach them to a specific chat when they're only relevant to that conversation. And it works both ways - you can now not only upload files, but also ask the agent to generate files for you, like PDFs and CSVs, right in the chat.

### Introducing ClarityQ Tokens (CQT) - a new way to meter and pay for your usage

ClarityQ Tokens (CQT) are a new usage and billing framework: transparent token metering, org-level budgets and top-ups and a real-time usage dashboard for admins. CQT gives you full visibility and control over how your team uses ClarityQ - and a more flexible way to pay for it. Curious whether CQT is right for your organization? Reach out to your ClarityQ contact and we'll walk you through it.

**July 21, 2026**

## A faster agent, Rules, and feedback that fixes itself

### A significantly faster agent

This release brings a major performance upgrade - a new fast agent architecture with a smarter harness under the hood. You'll feel the difference immediately: quicker answers, snappier follow-ups, same accuracy.

### Introducing Rules

Rules let you attach instructions directly to a metric, dimension, or other Context Layer item - and they're automatically loaded into the agent's context whenever that item is used. It's the most reliable way to make sure business logic like "always exclude test users from revenue" is applied every single time. Head to the Context Builder to create and manage your Rules.

### Feedback that fixes itself

Give feedback on any answer and ClarityQ will now diagnose what went wrong and let the agent correct itself - right there in the conversation. Better answers, and a smarter agent over time.

**June 25, 2026**

## Faster Context Builder and MCP Integrations

### Faster Context Builder

Significant performance improvements under the hood. You should notice shorter build times across the board.

### New - MCP Integrations!

You can now connect ClarityQ to external data sources through MCP. AppsFlyer, Notion, and any internal MCP server your company runs - all can now be queried in plain language, just like the rest of your data.

**June 11, 2026**

## Introducing Dashboards

Turn any question into a dashboard - built and maintained in plain language. Freely chat with the dashboard to dig deeper ("why is there a drop?") or make it public to share with anyone on the team.

**May 28, 2026**

## Visuals everywhere you work

Slack responses now include charts inline, and task results land in your inbox with the visual attached. Same view you get in Ask Anything - now everywhere ClarityQ reaches you.

**May 17, 2026**

## Self-serve Context Builder, Product Memory and ClarityQ in Slack

### Context Layer building, now self-serve

You can now build and edit anything in your Context Layer - including skills - entirely in natural language, optionally using files as input, with full versioning so you can review and roll back changes safely. Admins and Analyst-Contributors can open the new Context Builder page to try it.

### Product Memory is now yours to manage

Product Memory is the persistent knowledge ClarityQ keeps about your product, and it's loaded into the context of every session your team runs in this product. You can now see what's stored and how it shapes the agent's answers. Admins and Analyst-Contributors can also edit Product Memory as part of Context Layer versions - find it on the Memory page.

### Ask ClarityQ from Slack

You can now ask questions directly in Slack - mention @ClarityQ in any channel or DM, get answers in a thread, follow up, and click "Continue in ClarityQ" to jump into the full chat with visualizations in the web app. Ask your admin to reconnect the Slack integration with the new permissions to enable this.

### See what's synced at a glance

You can now see a "Synced" status on each Event Catalog item, so you can quickly tell which events, params and properties are already in your Semantic Catalog and which still need syncing. Open the Event Catalog to check.

**April 12, 2026**

## API keys

Admins can now create API keys to connect with ClarityQ programmatically. We'll continue expanding what's available through the API based on your needs. Check out the full documentation at [https://docs.clarityq.ai/api-reference](https://docs.clarityq.ai/api-reference.md).

**March 31, 2026**

## Automated Tasks and Slack delivery

### Automated Tasks

Schedule your Ask Anything questions to run automatically on a daily, weekly, or monthly basis. Create a task directly from a conversation - just ask the agent to schedule it, and it will craft the right prompt based on your chat. You can also define the prompt yourself from the Tasks tab on the Automations page. Each result includes a summary of the key takeaways and is delivered to your Results tab, email, or Slack. You and your teammates can continue the conversation on any result to explore it further.

### Slack Integration

Receive your scheduled task results directly in the Slack channels of your choice, and assign a different channel to each task. To get started, ask your ClarityQ workspace admin to connect ClarityQ to Slack from the Configuration page.

This release also includes improvements to the data warehouse setup process and various bug fixes.

**March 19, 2026**

## Skills replace Recipes, and a new Context Layer section

### Skills - Recipes, Evolved

What you knew as Recipes are now **Skills** - and they've gotten a major upgrade. All your existing recipes have been automatically migrated, so there's nothing you need to do.

Here's what changed:

* **Skills live in their own section** - they're no longer nested inside the Semantic Catalog, making them easier to find and manage
* **More powerful** - Skills can now be extended with references for deeper context, and come with built-in clarification questions and ToDo lists that guide the agent step by step
* **Invoke them on demand** - use / slash commands in chat or pick from the new Skills menu in the input area to run any skill proactively

### Refreshed Navigation - Context Layer

We've grouped all the components that shape your agent's knowledge under one unified section called **Context Layer**. This includes:

* **Catalogs** (formerly a standalone section)
* **Skills** in their own dedicated page
* **Memory** (moved from the user menu) - currently stores your personal memory, with product and organization memory coming soon

Everything the agent uses to understand your data and how to work with it - now in one place.

### Enhanced Chat Feedback

We've made it easier to share your thoughts:

* **New "Send Feedback" button** - share general feedback anytime, not just when something goes wrong
* **Add comments to positive feedback** - tell us what you loved when you give a thumbs up

### "How Much Time Did You Save?"

After productive chat sessions, you may see a quick banner asking how much time ClarityQ saved you. Your responses help us keep improving. It's lightweight, non-intrusive, and you can dismiss it anytime.

**March 5, 2026**

## User Management

We've introduced User Management to provide more control over access in ClarityQ.

Admins can now:

* **Invite users in the organization to ClarityQ**
* **Grant users access to specific products**
* **Assign and manage user roles for each product**

This makes it easier to manage teams and ensure the right people have the right level of access across ClarityQ.

**February 23, 2026**

## Long chats that never lose context, and Submit a Request

### Smarter Conversations That Never Lose Context

Long chats now handle their context limit gracefully. When a conversation gets long, you can "compact" it to keep going - either with an automatic summary or by telling ClarityQ what context matters most.

### Submit a Request

Got feedback, a request for catalog change, or an improvement idea? You can now submit requests directly from ClarityQ. Just click "Submit a Request" in the user menu, tell us what you need, and our team will get back to you.

### Smarter AI Under the Hood

Upgraded our underlying model to Claude Opus 4.6 for faster, more accurate responses.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.