> For the complete documentation index, see [llms.txt](https://docs.uptiq.ai/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.uptiq.ai/ai-workbench-qore/release-notes/2026/october-31.md).

# October 31

What's New in QORE

> *This October, one theme: control. Control over what goes live, what it costs, who can see it, and how well your agents perform.*
>
> *Build, test, publish, share, measure. This release connects every step of that journey, so what you build can reach production with checks along the way and no surprises after. Here's what's new.*

{% columns %}
{% column width="33.33333333333333%" %}

## <i class="fa-book-open">:book-open:</i> Knowledge & Govern Graphs

*Answers from knowledge you trust.*
{% endcolumn %}

{% column width="66.66666666666667%" %}
![Graph Explorer showing govern graph nodes and relationships](https://1326225582-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2F0qmgQjJ5aArDTj2ACFHG%2Fuploads%2Fgit-blob-4e33d706171a4d74d98d2a953bf8e9e13536b4b4%2Fgraph-explorer.png?alt=media)

* Build govern graphs in your own account. See [Govern Graph](https://docs.uptiq.ai/ai-workbench-qore/control-center/govern/govern-graph.md).
* Edit graph nodes in Graph Explorer. Every edit is saved as a new graph version. See [Edit a node](https://docs.uptiq.ai/ai-workbench-qore/control-center/govern/govern-graph.md#node-editing).
* Agents get a ready-made Cypher query tool (query language for Neo4j), and large answers are no longer cut off midway. See [Skills & Capabilities](https://docs.uptiq.ai/ai-workbench-qore/console/agent-builder/build/skills-and-capabilities.md).
* Graphs travel with agent and knowledge packages on export and import.
* Import files straight from an S3 or Google Cloud Storage link.
* Revert a knowledge version, preview older ones, and see files added after sharing. See [Sync and versions](https://docs.uptiq.ai/ai-workbench-qore/console/knowledge/sync-and-versions.md).
  {% endcolumn %}
  {% endcolumns %}

{% columns %}
{% column width="33.33333333333333%" %}

## <i class="fa-scale-balanced">:scale-balanced:</i> Evaluation & Observability

*See what your agents are doing, and how well.*
{% endcolumn %}

{% column width="66.66666666666667%" %}
![Regression Eval run results with passed and failed test cases](https://1326225582-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2F0qmgQjJ5aArDTj2ACFHG%2Fuploads%2Fgit-blob-298584db3f049c2aab29efc4c86c78fa151c2581%2Fregression-run.png?alt=media)

> **Two kinds of evals, two moments:**
>
> * **Regression Eval** runs *before* an agent is published. It tests the agent against your golden data, so you know it meets the bar before it goes live.
> * **Online Eval** runs on agents already *in production*. It scores real conversations as they happen.

#### Regression Eval

* "Evals" is now **Regression Eval**, with a cleaner layout. See [Regression Evals](https://docs.uptiq.ai/ai-workbench-qore/console/agent-builder/test/regression-evals.md).
* **Bring your own test data.** Upload your regression test data, so you can test agents against the cases that matter to your business.

#### Online Eval

* **Checklists are now an evaluation method.** Add a checklist in Evals and run it against your golden data sample. Each sample is checked item by item, so you can see exactly which requirements an agent meets or misses. See [Checklist](https://docs.uptiq.ai/ai-workbench-qore/control-center/evals/evaluators.md#checklist).
* **Agents can be evaluators.** Use an agent as an online evaluation method to score live production conversations. Every score is recorded per judge. See [Choose a framework](https://docs.uptiq.ai/ai-workbench-qore/control-center/evals/evaluators.md#frameworks).

![Online Eval Runs with scores and verdicts](https://1326225582-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2F0qmgQjJ5aArDTj2ACFHG%2Fuploads%2Fgit-blob-bea1d65242f583c48078f0f9551b417e16ca3763%2Fonline-runs.png?alt=media)

#### Runs & Traces

* Every conversation, run, and trace gets its own page. See [Runs & Traces](https://docs.uptiq.ai/ai-workbench-qore/control-center/observe/runs-and-traces.md).
* A new **Project** column helps tell apart agents with the same name.
* Tool names show on tool-call steps.
* Trace summaries can be generated on demand.
* Token totals and span counts now line up.
* A new observability link sits in the Operator's conversation history.
  {% endcolumn %}
  {% endcolumns %}

{% columns %}
{% column width="33.33333333333333%" %}

## <i class="fa-rocket">:rocket:</i> Publish, Version, Review

*Ship it. Review it. Roll it back if you need to.*
{% endcolumn %}

{% column width="66.66666666666667%" %}
*.*

* **Publish anything, with confidence.** Agents, apps, design systems, flows, and knowledge now share one [publish and review process](https://docs.uptiq.ai/ai-workbench-qore/control-center/govern/publish-review.md).
* Design systems and flows can require approval before going live. Don't need it? Switch it off per action.
* Every publish request links straight to the asset under review.

#### Agent versioning

* Agents have saved versions: save a version, run the one you trust, and keep building the next. See [Runtime & Versioning](https://docs.uptiq.ai/ai-workbench-qore/console/agent-builder/build/runtime-and-versioning.md#which-agent-version-runs).
* Agents, the Agent SDK, and flows can run a specific version. See [Add a flow as a skill](https://docs.uptiq.ai/ai-workbench-qore/console/agent-builder/build/adding-a-skill-to-the-agent.md#add-a-flow-as-a-skill).

#### More control over what ships

* Went the wrong way in App Builder? Ask it in chat to restore an earlier version. See [Versions and publishing](https://docs.uptiq.ai/ai-workbench-qore/console/app-builder/app-workspace.md#versions-and-publishing).
* Knowledge can't be published while it's still building, so no half-finished releases.
  {% endcolumn %}
  {% endcolumns %}

{% columns %}
{% column width="33.33333333333333%" %}

## <i class="fa-coins">:coins:</i> Cost Control

*Budgets you can actually set.*
{% endcolumn %}

{% column width="66.66666666666667%" %}
![Credit Usage dialog showing daily and weekly limits](https://1326225582-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2F0qmgQjJ5aArDTj2ACFHG%2Fuploads%2Fgit-blob-b385264b7a7cf7c5ef945d05f122bd8d5f65c0b3%2Fcredit-usage-limits.png?alt=media)

* **No more surprise bills.** Cap agents by tokens, dollars, or per conversation. See [Credit Usage Limit guardrail](https://docs.uptiq.ai/ai-workbench-qore/console/agent-builder/deploy/guardrails/credit-usage.md).
* New accounts need a usage limit from day one, and the last hard limit can't be deleted. See [Usage Limits](https://docs.uptiq.ai/ai-workbench-qore/control-center/cost-control/usage-limits.md).
* [Credit Usage](https://docs.uptiq.ai/ai-workbench-qore/control-center/cost-control/credit-usage.md) and Usage Limits pages have a fresh redesign.
* Operator conversations show what each reply and each conversation cost.
* Per-turn token usage is available through the public API and the App Builder MCP, with a full breakdown in Runs and Traces. See [Read a turn's token usage](https://docs.uptiq.ai/ai-workbench-qore/console/app-builder/creating-and-importing-apps.md#read-a-turns-token-usage).
* Admins can sort accounts by credit limit.
  {% endcolumn %}
  {% endcolumns %}

{% columns %}
{% column width="33.33333333333333%" %}

## <i class="fa-shield-halved">:shield-halved:</i> Access, Roles & Security

*The right people, the right access.*
{% endcolumn %}

{% column width="66.66666666666667%" %}
![Share dialog showing people with access and linked assets](https://1326225582-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2F0qmgQjJ5aArDTj2ACFHG%2Fuploads%2Fgit-blob-032864c703d7b53972b4b6d53e6066275c68eb4b%2Fshare-dialog.png?alt=media)

#### Roles and sharing

* **Teams get roles.** Assign roles to a team once, and manage [teams](https://docs.uptiq.ai/ai-workbench-qore/control-center/admin/teams.md) from a single dialog.
* **Share without the chaos.** Give teammates Viewer or Editor access to what you've built. [Apps, datasources, design systems, knowledge, and flows](https://docs.uptiq.ai/ai-workbench-qore/control-center/admin/sharing.md) can be shared with users and teams. Agents can be shared with service accounts.
* **See what's linked.** The share window shows linked assets, and editors can browse them by version.

#### Keys and sign-in

* **Safer API keys.** Permissions are editable, and a [key](https://docs.uptiq.ai/ai-workbench-qore/control-center/admin/api-keys.md) can never do more than its creator's role allows.
* **Stronger sign-in.** Super admins can reset a user's MFA, and users can turn on MFA from their profile. Password rules are enforced on the server, with a reuse limit per organization. See [Two-factor authentication](https://docs.uptiq.ai/ai-workbench-qore/control-center/admin/password-sign-in.md#two-factor-authentication).
* **Privacy fix.** For SharePoint or OneDrive connection.
* Security patches landed across the platform, from the app generator to the backend.
  {% endcolumn %}
  {% endcolumns %}

{% columns %}
{% column width="33.33333333333333%" %}

## <i class="fa-robot">:robot:</i> Agent Builder & Operator

*Build smarter agents, and let them work with others.*
{% endcolumn %}

{% column width="66.66666666666667%" %}

#### Agents working with others

* **Agents can talk to partners.** Partner agents can discover and call your agents over the Agent2Agent (A2A) protocol. See [Access & Security](https://docs.uptiq.ai/ai-workbench-qore/console/agent-builder/deploy/authentication.md#a2a).
* **Copilot or Auto.** Operator and the headless trigger both support the two modes. See [Auto and Copilot](https://docs.uptiq.ai/ai-workbench-qore/console/console-operator.md#auto-and-copilot).
* **Smarter sub-agents.** The parent agent chooses what context to hand over. Sub-agent UI renders properly instead of raw JSON.

![Agent Config Access & Security with the A2A tab](https://1326225582-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2F0qmgQjJ5aArDTj2ACFHG%2Fuploads%2Fgit-blob-464bb703e5929692a9251ceb6fca470831dd2b4c%2Fa2a-tab.png?alt=media)

#### Building and configuring

* **Longer instructions.** The 3,000-character cap is gone, and a warning appears above 5,000. See [Instructions](https://docs.uptiq.ai/ai-workbench-qore/console/agent-builder/build/instructions.md).
* **Integration Hub.** Choose which methods to enable when you connect. See [Use integrations from an MCP client](https://docs.uptiq.ai/ai-workbench-qore/console/integrations-hub.md#use-integrations-from-an-mcp-client).
* **Tidier config.** [Secrets & Variables](https://docs.uptiq.ai/ai-workbench-qore/console/agent-builder/build/secrets-and-variables.md) are split into Secrets, Variables, and Built-in tabs. Access & Security becomes Access Keys and Allowed Domains. FAQs, memories, and learning cases show in tables.
* Execution plans are revamped, the playground loads faster, and Agent Builder no longer asks for approval twice.
  {% endcolumn %}
  {% endcolumns %}

{% columns %}
{% column width="33.33333333333333%" %}

## <i class="fa-swatchbook">:swatchbook:</i> App Builder

*More ways to bring your work in.*
{% endcolumn %}

{% column width="66.66666666666667%" %}

* Put published apps built elsewhere on your own custom domain. See [Domains](https://docs.uptiq.ai/ai-workbench-qore/console/app-builder/build/app-config.md#domains).
* Messy naming and styling in frontend code get caught before it's saved.
  {% endcolumn %}
  {% endcolumns %}

{% columns %}
{% column width="33.33333333333333%" %}

### <i class="fa-sparkles">:sparkles:</i> A Fresher Look

{% endcolumn %}

{% column width="66.66666666666667%" %}

* Control Center, Org Settings, and the Store now share one layout.
* Dark mode and table row actions are consistent across the workbench.
  {% endcolumn %}
  {% endcolumns %}

{% columns %}
{% column width="33.33333333333333%" %}

### <i class="fa-bug">:bug:</i> Fixes & Stability

{% endcolumn %}

{% column width="66.66666666666667%" %}

* Agent import on Azure no longer crashes the server.
* Redis stays steady when many agents run at once.
* App Builder recovers after you stop a generation.
* Error messages are clearer. Provider-unavailable and Gemini failure reasons now show up.
* A failing agent turn ends the run instead of repeating.
* Abandoned LLM calls are cancelled, so you don't pay for work no one is waiting for.
* Billing app ID passes correctly between services (hotfix).
  {% endcolumn %}
  {% endcolumns %}

{% hint style="success" icon="rocket-launch" %}
There's more on the way, and we're already working on the next set of capabilities.

*Try the new features, tell us what works and what doesn't, and keep building. We're listening.*
{% endhint %}


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation by asking a question.

Perform an HTTP GET request on the following URL with the `ask` and `goal` query parameters:

```
GET https://docs.uptiq.ai/ai-workbench-qore/release-notes/2026/october-31.md?ask=<question>&goal=<user_goal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is what the user is ultimately trying to achieve, the reason they need the answer. Sharing it helps GitBook give you a better, more relevant answer. A goal is most helpful when it describes the outcome the user wants rather than restating the question. For example, with `ask=how do I create an API token`, a goal like `automate deployments from our CI pipeline` lets GitBook tailor the answer to that use case.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.
