Open edX Verawood: What's New and What Breaks

If you're still running Ulmo, your release-support clock has already run out. Open edX Verawood, the 22nd community release, shipped on July 30, 2026, and with it every release before it fell out of community support. That includes Ulmo, which many production sites had only finished upgrading to a few months earlier.

This article is the operator-side briefing on Verawood: the verified feature list, the deprecations you need to audit for, and a realistic 4–8 week upgrade timeline. We've shipped Open edX deployments at Cubite since 2013, across Sumac, Teak, and Ulmo. This is the same brief we'd hand our own engineering team.

The 30-Second Brief on Verawood

Open edX Verawood is the 22nd community release of the Open edX platform, released on July 30, 2026 per the official Verawood release notes. It was originally scheduled for June 9 on the Open edX release schedule, then moved to June 23 by the Build-Test-Release working group. The release branch was cut at the end of April 2026, and the release notes cover changes from October 30, 2025 to April 9, 2026.

In our reading, Verawood is about platform maturity and operational stability: fewer flashy features, more long-term groundwork. Headline items:

• AI Extension Framework: a standardized plugin for AI features (it also runs on Teak and Ulmo; Verawood is the first release whose release notes cover it)
• Instructor Dashboard overhaul: the new React/Paragon dashboard replaces the Django-template version and is on by default
• Granular team management: an opt-in, role-based permission system for Studio course authoring (stackable program pathways are planned for Willow)
• LTI + Content Library hardening for reliable block-level reusability
• Frontend foundation: first release to support frontend-base (OEP-65), alongside the Paragon design-token theming (tutor-contrib-paragon) available since Teak
• Tutor v22 support (released August 2026), which moves edx-platform to Python 3.12 and requires a Meilisearch reindex

The Open edX Release Calendar, Quickly Reframed

Open edX runs a six-month release cycle. Named versions are scheduled for June 9 and December 9, though recent releases have slipped: Ulmo shipped in January 2026 and Verawood on July 30, 2026. Per the Open edX release schedule wiki:

Practical read: if you're on Teak or earlier, you're not just behind on features: you're running code with no upstream CVE coverage. If you are on Ulmo, its official support ended on July 30, 2026. The August and September 2026 openedx-platform security fixes were still added to Ulmo, but don't plan around that continuing.

(Still weighing the platform itself rather than the upgrade? See our Open edX vs Moodle comparison before going further.)

What's Actually in Verawood — The Feature Deep Dive

Below are the Open edX Verawood features worth knowing as an operator. Claims link to a primary source where one exists: docs.openedx.org, the openedx-ai-extensions repo, or the relevant GitHub deprecation issue.

The AI Extension Framework

A notable strategic shift around Verawood is the Open edX AI Extension Framework. It's a platform plugin (openedx-ai-extensions) that introduces a standardized integration path for AI tools instead of one-off bolt-ons. The plugin also runs on Teak and Ulmo; Verawood is the first release whose official release notes cover it.

Concretely, the framework provides:

• A LiteLLM-backed provider layer: supports OpenAI, Anthropic, and local models out of the box
• Tool primitives already shipped in the demo fixtures: content summarization, flashcard generation, chat with function calling, streaming chat, and an educator assistant
• MFE plugin slots in both the Learning MFE and the Authoring MFE for surfacing AI experiences to learners and authors
• Aspects event analytics so every AI workflow is observable
• Configuration-driven behavior: most setup happens in Django admin at /admin/openedx_ai_extensions/, not in code

The caveat that matters: per the project README, the framework is currently experimental and not yet recommended for production. The Verawood release notes do describe working features, including an in-unit AI assistant for learners and AI-drafted multiple-choice questions in Studio, but the README warning still applies. Operators who set up the plugin pipeline now will be ready when first-party AI tools mature in subsequent releases — that's the win for early adopters.

The Instructor Dashboard MFE Migration

The Instructor Dashboard has been a server-rendered Django-template island in an otherwise MFE-first platform for years. Per GitHub issue #37214, that was the plan, and in Verawood it landed: the new Instructor Dashboard is the default experience. That ticket has since been closed as a duplicate; the open deprecation ticket, issue #38432, lists Xylon as the earliest named release with breaking changes.

The new dashboard lives in frontend-app-instructor-dashboard (originally frontend-app-instruct) and is built on React, TypeScript, React Query, and Paragon. Ten Django template files are slated for eventual removal, covering: course info, membership, cohort management, student administration, data downloads, special exams, certificates, and more.

If your deployment includes custom instructor flows or template overrides (and many multi-tenant or higher-ed deployments do), this is one of the highest-impact deprecations in Verawood. Plan for a port, not a copy-paste. In Verawood the new dashboard is on by default, and sections injected through the legacy InstructorDashboardRenderStarted filter no longer appear. The instructor.legacy_instructor_dashboard waffle flag brings the old dashboard back, but only as a stopgap: the legacy dashboard is deprecated and expected to be removed in a future release (the open deprecation ticket lists Xylon as the earliest named release with breaking changes). Verawood is where the migration starts, not where it ends.

Stackable Program Pathways (Planned for Willow)

An earlier version of this article listed stackable pathways for programs (micro-credentials feeding into degrees, prerequisite chains, sequenced certifications) as a Verawood feature. They did not ship in Verawood: the official Verawood overview lists Modular Pathway Support under "Upcoming in Willow" and says that for Willow the platform will support stackable pathways for programs. Until then, expressing real curriculum architecture in Open edX usually still means external workflow tools or custom XBlocks.

For higher-ed and workforce-credential operators, this is the feature to watch in Willow. For pure self-paced course catalogs, it's mostly invisible, and that's fine.

LTI + Content Library Hardening

Verawood builds on Ulmo's LTI Advantage Complete certification and reusable LTI configurations with two operator-grade improvements: reliable LTI block duplication (which has been historically brittle) and a simpler, tabbed LTI setup in Studio. If you've ever lost a day debugging an LTI tool that worked in the source course but not in the duplicate, this is for you.

Frontend and Tutor

Verawood is the first release that supports frontend-base, the unified modern frontend foundation specified in OEP-65. It ships four frontend-base apps: the new Instructor Dashboard and notifications are on by default, while the Authn and Learner Dashboard apps are opt-in. Separately, the tutor-contrib-paragon plugin (available since Teak) compiles Paragon design tokens into MFE themes through Tutor; frontend-base apps follow their own theming docs, so existing MFE branding will likely need porting.

On the operations side, Tutor users move to Tutor v22 (released August 2026), which requires Python 3.10+, moves edx-platform to Python 3.12, and needs a Meilisearch reindex during the upgrade. The longer-term payoff is frontend-base: the community expects all MFEs to be converted to frontend-base apps by Xylon (June 2027). Translation: this release pays you back at next upgrade, not this one.

What Breaks, Deprecates, and Disappears

The full Verawood breaking-change list is in the Developer & Operator Release Notes, which are the source of truth. The items we'd audit first:

• Legacy Instructor Dashboard: the new frontend-app-instructor-dashboard is the default, the Django-template dashboard is deprecated (issue #38432), and sections injected through the legacy InstructorDashboardRenderStarted filter no longer appear. Audit any template overrides or instructor-tool customizations.
• Legacy frontend patterns predating frontend-base/OEP-65 — anything still hooking the pre-frontend-base scaffolding is on borrowed time.
• Paragon v22 patterns — if your theme code predates v23 design tokens, plan a port.
• Plugins not validated against the Ulmo-cycle Tutor changes: pip → uv and the Tutor Deck transition can break community plugins that were never updated for them. Anything you haven't smoke-tested under Tutor v22 is suspect.
• Aspects Analytics v4.0.0: a breaking upgrade that moves Aspects to Python 3.12 and Superset 6.0. Review it before touching any analytics-pipeline customizations.

A plain Ulmo deployment with no forks and standard plugins will see a fairly mild Verawood. A heavily forked, multi-tenant, custom-instructor-dashboard deployment will not. Both deserve the audit; only one can safely skip the phased plan in the FAQ below.

Why You Can't Just Wait

The cleanest argument for prioritizing your Open edX upgrade in 2026 is the support model itself. Per the Open edX named release docs, the latest release is the only supported release, and the community states it does not currently have the ability to support more than one release at a time.

Two compounding factors most operators underweight:

Operational debt compounds non-linearly. Each release you skip multiplies migration squashes, plugin breakages, and theme rewrites. The Open edX team periodically squashes Django migrations across versions, and selecting the right migration entry point is the difference between a clean upgrade and a corrupted database — as Raccoon Gang's upgrade guide puts it bluntly. Three skipped releases isn't 3× harder. It's closer to 6×.

EOL software is becoming a documented legal liability in the EU. The Cyber Resilience Act (CRA), whose reporting obligations have applied since 11 September 2026 and whose main obligations apply from 11 December 2027, makes vendors accountable for vulnerability handling in every component they ship — including end-of-life upstream open source. Per HeroDevs' analysis of the CRA, the regulation makes EOL software in your product stack a documented legal liability: you own the vulnerability-handling obligations for every component you ship, regardless of whether that component still has an upstream maintainer. If you ship software products to EU customers, check your CRA obligations now.

The defensible "wait for .1" position has expired: Verawood.1 has been out since July 30, 2026. From Teak or earlier, every month of delay buys you more debt, not safety.

Why Most Teams Bring in a Partner (And What "Reliable" Looks Like)

Doing the Open edX upgrade in-house is viable. We've worked alongside teams who do it well. The shared traits are predictable: they upgrade more than once every 18 months, they have full-time staff fluent in Tutor / Django / MFEs, and their fork is small or vanilla.

If even one of those isn't true, the math usually tips toward a partner. A short DIY-vs-partner checklist:

What to look for in a partner:

• Verifiable Open edX credentials: a listing in the Open edX provider directory, public code contributions, or client references
• A multi-release track record: ask specifically whether they shipped Teak → Ulmo and Ulmo → Verawood upgrades and what broke
• Transparent methodology and rollback policy — not just "we'll handle it"
• Real engineering depth — Tutor plugins, custom XBlocks, MFE work, not just managed hosting

Cubite has been engineering Open edX deployments since 2013. We offer dedicated Open edX hosting from $1,260/month, custom platform builds, and Verawood upgrade assessments. The upgrade plan in this article is the same one our engineering team runs.

FAQ

What is Open edX Verawood?

Open edX Verawood is the 22nd community release of the Open edX platform, released on July 30, 2026. It focuses on platform maturity and modernization: a new MFE-based Instructor Dashboard (on by default), the AI Extension Framework, first support for frontend-base, Content Library and LTI improvements, and opt-in role-based permissions for Studio. Its release also ended community support for Ulmo.

Verawood's release notes cover changes from October 30, 2025 to April 9, 2026, and the release branch was cut at the end of April 2026. In our reading the release favors operational stability over flashy features: much of the work is groundwork that pays back at the next upgrade, such as first-release frontend-base support, the new Instructor Dashboard, and LTI blocks that can now be duplicated. See the official Verawood release notes for the canonical feature list.

When is Open edX Verawood released?

Open edX Verawood was released on July 30, 2026, after its release branch was cut at the end of April 2026. It was originally scheduled for June 9, moved to June 23, and then shipped in late July. It succeeds Ulmo (January 2026) on Open edX's six-month release cadence. With Verawood's release, Ulmo lost community support, so if you're still on Ulmo, now is the time to plan and stage your upgrade.

The cadence targets June 9 and December 9, but dates slip: Verawood shipped about seven weeks late. Willow, the next release after Verawood, has been pushed back too: the Build-Test-Release working group plans to cut it in the first week of January 2027 and is targeting a mid-to-late February 2027 release.

What's new in Open edX Verawood?

Verawood ships a new MFE-based Instructor Dashboard built with React and Paragon (on by default), the AI Extension Framework (LiteLLM-powered, supporting OpenAI, Anthropic, and self-hosted models), first support for frontend-base, duplicable LTI blocks with a simpler LTI setup, Content Library improvements, a new Studio sidebar, in-platform taxonomy tagging, and opt-in role-based permissions for Studio. Stackable program pathways are planned for Willow.

The AI framework is the strategic headline; the Instructor Dashboard MFE migration is the practical one. If your deployment customizes the dashboard, that's where most of your Verawood engineering work will live.

Do I have to upgrade from Ulmo to Verawood?

Strictly, no — Open edX is open source. Practically, yes: only the latest release is officially supported. When Verawood shipped on July 30, 2026, Ulmo joined Teak and Sumac in the unsupported tier. The August and September 2026 security fixes were still added to Ulmo, but that is not guaranteed, and any CVE that isn't backported becomes your team's problem. For most operators, planned upgrade beats unplanned incident response.

The EU Cyber Resilience Act adds a legal dimension: shipping EOL software to EU users will carry documented vendor obligations from December 2027. Combined with the operational-debt math (each skipped release compounds upgrade complexity), the "wait" position has run out for Ulmo sites now that Verawood.1 is out.

How long does an Open edX upgrade take?

A planned Ulmo to Verawood upgrade typically takes 4–8 weeks for a moderately customized instance: 1 week audit, 1–2 weeks fork reconciliation and plugin compatibility, 1–2 weeks staging build, 1 week dry-run migration with squashed Django migrations, and a weekend cutover with rollback gate. Heavily forked or multi-tenant deployments take longer.

Each release you've skipped roughly compounds the timeline. A Teak → Verawood upgrade isn't 2× harder than Ulmo → Verawood — it's more like 3–4× because of migration squashes, plugin compatibility gaps, and theme version drift.

Should I use a managed provider to upgrade Open edX?

If your team upgrades Open edX more often than once every 18 months and has staff fluent in Tutor, Django migrations, MFE builds, and your specific fork, in-house is viable. If not, an Open edX migration service provider almost always ships faster and safer — once you account for opportunity cost of internal engineering time.

The clearest signal: if you're more than one release behind (Teak or earlier) and you have any production customization, the risk profile strongly favors a partner.

Will my plugins break in Verawood?

Don't assume a plugin is Verawood-ready: check each one for a release that supports Tutor v22. The breakages cluster in two places: plugins that haven't been updated through the Ulmo-cycle Tutor changes (pip → uv, Tutor Deck) and plugins that hook the legacy Instructor Dashboard templates. Both surface during the plugin-compatibility phase of the upgrade timeline above.

The forensic approach: for each plugin, check the latest release date, whether it has a Verawood-tagged version, and whether its issue tracker has open compatibility threads.

Is Open edX Verawood backward compatible with my current Tutor version?

Not always. Each Open edX named release pins to a major Tutor version, and Verawood continues the modernization that started in Ulmo (uv-based builds and Tutor Deck in Tutor v21, plus tutor-contrib-paragon theming since Teak). Tutor added Verawood support in Tutor v22, which requires Python 3.10+ and a Meilisearch reindex. Plan to bump Tutor as part of the upgrade, and validate every Tutor plugin against the new version before staging cutover.

The Tutor releases page is the canonical source; lock the Tutor version in your deployment manifest rather than tracking latest.

The Bottom Line

Open edX Verawood is a consequential platform release. The AI Extension Framework changes how AI integrations are built. The Instructor Dashboard MFE migration changes what every operator with custom instructor flows has to maintain. Frontend-base changes how MFEs will be built and themed from here on.

But the most actionable fact about Verawood isn't a feature. It's the date: July 30, 2026. When Verawood shipped, Ulmo lost community support, Teak and Sumac sank deeper into unsupported territory, and the operational + legal cost of staying behind started compounding faster.

If you're on Ulmo or earlier, the right move now is the pre-upgrade audit, followed quickly by the upgrade itself. Know your forks. Know your plugins. Know what your instructor dashboard customizations are about to cost you. Then plan the migration window with the phased timeline in the FAQ above.