Technical Oversight Committee Meeting Notes
Announcements and Community Updates [10:34:19 - 10:53:24]
Continued TOC Participation
Anant Agarwal has transitioned to senior advisor at 2U, but his involvement with the Open edX project and the TOC would continue.
Conference Recap and Future Planning
The recent conference, held in May and co-hosted by WGU, was reported as a successful event, though slightly smaller than previous years. A significant challenge identified was difficulty obtaining visas for international attendees, with one organization experiencing nearly complete visa rejections except for attendees traveling from France. The community emerged energized from the event, with several new initiatives already gaining momentum.
Planning for next year’s conference was underway, but the destination has not yet been finalized.
Digital Public Infrastructure Focus
A strategic focus area was described around developing the ecosystem of organizations using the platform, particularly abroad and in the context of digital public infrastructure. Based on existing proof points and case studies, confidence was expressed that the Open edX platform already has product-market fit in this space. Business development-adjacent activities were being pursued to drive adoption and awareness. Examples cited included work in Ethiopia and the NELC initiative in the Kingdom of Saudi Arabia. Attendance at various events had generated numerous conversations and potential adoption opportunities.
Recently Engaged Platforms
Several recently engaged organizations were highlighted. The WHO Academy was deepening its engagement, hiring Open edX technical resources in Kuala Lumpur with a team of around 10 people working on the platform, presenting an opportunity to potentially bring them on as a mission-aligned organization over time. A case study was shared for the Clemson University Center for Workforce Development, a skills-forward platform based on the Open edX platform focused on specific skill-oriented job opportunities.
Open Source SIS Investigation [10:41:19 - 10:47:47]
A question was posed to the group regarding familiarity with open source Student Information Systems, a need frequently heard from adopting organizations. Two specific projects were mentioned: Eagle Apps (created by Boston College, with a small consortium of users and partnership with Dexterra for API standards recommendations) and Open Campus (an EU-based open source SIS). A meeting with the Boston College team was scheduled for later that day.
Strong consensus emerged that building the Open edX platform into an SIS would not be advisable, as content delivery systems and Student Information Systems are quite distinct and would represent an entirely different, highly complex codebase and project. It was noted that after completing a campus project, the decision was made not to integrate directly with SIS systems because universities have so many different systems, though sometimes multiple universities share the same SIS.
The recommended approach was to have good APIs enabling three core functions: creating courses, creating groups, and sending grades back to the Student Information System. These three capabilities were considered sufficient to collaborate with any student administration system. The observation was made that while an SIS initially appears simple, universities each want their own personalization, resulting in complex systems that are similar but different, confirming it as a separate project of entirely different nature.
The clarification was made that there was no plan to build an SIS alternative, but rather a desire to be able to recommend something managed by another foundation that would be a configurable platform rather than a product. A proposed next step was to conduct research or a survey of available open source SIS options, documenting which universities have adopted each, and producing a white paper with investigation into how to link these to the Open edX platform. WGU was noted as using an on-premise version of Banner as its SIS, with recognition that changing SIS systems is akin to “open-heart surgery.”
End-to-End Testing Collaboration [10:47:47 - 10:53:24]
Coming out of the conference, significant community energy emerged around collaborating on end-to-end testing, with strong interest in using Playwright (Microsoft’s framework) rather than Cypress, which had been used for some existing tests. The developers driving this work find Playwright to be the superior tool from a DevX, developer experience, perspective.
A new collaborative working space was established for community members from WGU, MIT, and Raccoon Gang to harness their energy toward building something useful for the community. The initial goal was to build out capabilities to replace current manual testing around the release cycle, thereby speeding up releases.
The vision described included a base set of smoke tests and end-to-end tests that everyone deploying the platform would run before releasing to production, with the ability for organizations to layer on custom tests specific to their environments (such as AI chatbots or specialized features). The core test suites would reduce costs for functionality everyone needs, like enrollment, registration, and login. 2U was noted as also using Playwright with smoke tests and regression tests derived from the BTR test suite, though these were currently in a private repository. A conference talk on the testing system design was referenced as available on the Open edX YouTube channel.
Frank Nagle Talk Recommendation
A recommendation was made to view Frank Nagle’s plenary talk, which presented research from an economics background into the value of open source. The analysis explored how open source is pivotal in creating value for businesses, how contribution creates a dividend for contributors, and how healthy projects with active contribution generate more value for both the central project and all participants—an approach starting neither from an ideological free-software perspective nor a purely marketing perspective.
REST API Standardization Project [10:53:24 - 11:06:45]
Project Motivation and Approach
A presentation was made on REST API standardization work, with implementation being led by Edly and architectural guidance and support provided separately. The motivation stemmed from 13 years of accumulated REST APIs and pseudo-REST HTTP endpoints that behave inconsistently. As the platform moves toward becoming more of a platform, well-defined REST endpoints for third-party integration were identified as increasingly valuable.
The project structure involved several phases:
Phase 1 (Standards Definition): Reviewing existing APIs and defining desired standards including authentication systems, valid response types, pagination, versioning, and documentation. This work went through the Architectural Decision Record (ADR) process, documenting standards in the Open edX Platform repository with community review.
Phase 2 (Implementation): Applying standards to core representative REST endpoints so they behave consistently—pagination, filtering, and search working the same way across endpoints. Documentation improvements would produce a more reliable OpenAPI spec.
Phase 3 (SDK Generation): Taking the spec and generating Python and JavaScript SDKs so third-party systems can integrate with the Open edX platform via backend Python systems, JavaScript systems, new front ends, or existing MFEs.
Progress Status
Approximately 14 ADRs had been completed and mostly merged, with implementation beginning to merge. One change to an OEP around standardization was in progress. New standardized API endpoints were shipping with version bumps to avoid breaking existing APIs. Existing APIs would remain, new ones would be introduced, internal tooling would migrate to the new ones, and old less-reliable APIs would go through the standard deprecation process. A published new OpenAPI spec with better documentation was expected within about a month.
ADR and OEP Clarification
For clarity, the distinction between the two document types was explained. Architectural Decision Records represent smaller decisions closer to code, while Open edX Proposals (OEPs) represent high-level decisions relevant to the entire Open edX codebase, defining high-level processes and approaches. The RBAC project work follows the high-level authorization/authentication approach, and the RBAC OEP required modification to clarify interaction with the new REST API standards. Django REST Framework was identified as the industry-standard framework used for most APIs over the past five years, with the ADRs defining consistent usage patterns meeting performance, security, and discoverability characteristics.
AI Acceleration and MCP Discussion
Discussion covered how a discoverable specification and SDK makes APIs more readily available for vibe coding and building AI applications on top of the platform’s APIs. AI coding tools had helped the team dive deep into corner cases and poorly documented API areas to make better decisions about deprecating versus replacing versus changing APIs.
A significant discussion emerged regarding Model Context Protocol (MCP) servers. A community request to build an MCP server into the platform was pushed back on, primarily because generating MCP servers based on REST API specs is fairly standard, avoiding the need to maintain both an API layer and an MCP layer. The position was that improving documentation and API consistency would allow anyone to generate MCP layers, just as Python or JavaScript SDKs are generated. Sharing the API spec tends to be sufficient to let models interact with third-party systems without building a full MCP server.
A counterpoint acknowledged that while sharing API specs works for one-off interactions, persistent processes (agents) would benefit from a smoother integration surface, though it was agreed the jury remains out on MCP versus skills versus plain REST. MCP was noted as receiving more pushback recently, making it sensible to wait before committing to maintaining more code. It was also noted that the AI experimentation framework (now in beta availability) explicitly chose function passing rather than MCP for embedded agents to access platform APIs, and this decision had not been regretted, having accomplished all intended use cases.
Campus Agora Initiative [11:07:12 - 11:16:51]
An update was provided on the Campus Agora initiative, conceived after the disappearance of Miríada X left many universities with courses lacking a platform, and following the transition where support for Spanish-speaking courses stepped back significantly. The idea emerged that a MOOC platform for the Hispanic market on the Open edX platform would be valuable.
The initiative was developed in collaboration with MetaRed (a branch of Universia, part of the social action of Banco Santander), which helps over 1,700 universities with technology. Banco Santander was receptive to the proposal, and the Open edX platform was presented as the best available solution. The bank found the idea compelling enough that they invited Santander Open Academy, which also adopted the Open edX platform for its initiative.
The primary challenge involved legal departments of the bank and universities. The agreement was signed in June, starting with four universities (Autónoma Madrid, Carlos III Madrid, Politécnica Madrid, and Politécnica Valencia), with Galileo awaiting to join. The initiative was opened to the entire MetaRed community, announced at conferences in Lisbon and Costa Rica, with additional universities now joining and first courses starting in September.
The financial model would start with free certificates for awareness, then seek self-sustainability. The structure would be a consortium ruled by a general assembly of universities together with the bank, including a quality committee to ensure high-quality courses and avoid past quality issues. Interest had been expressed from Portugal and 10 Latin American universities were beginning legal processes to join. A vision was described where universities in the consortium could both create courses and gain access to use other universities’ courses.
University LMS as Blackboard/Moodle Alternative [11:16:51 - 11:49:03]
Progress and Current Blockers
An update was provided on progress toward positioning the Open edX platform as an alternative to Blackboard and Moodle. Pilots were being launched or planned in several Iraqi universities and other Middle Eastern universities, targeting a specific locked-in feature set. Two remaining items requiring finalization were identified: trademark/ownership and funding.
Such an effort would create a new “product” that is separate and distinct from the Open edX core which would remain slim and focused on supporting many different modes of learning online, but in a flexible and general way.
Trademark and Branding Strategy
A trademark registration request had been submitted with several suggestions provided. The recommendation was to create a distinct brand incorporating the Open edX name but focused on universities—candidates mentioned included variations like “Open edX Campus” or similar university-focused naming. A distinct brand would make marketing the new product easier, as it would essentially be an Open edX distribution installed at universities that looks familiar to the LMS systems being replaced. Strong support was expressed for creating a separate brand to distinguish a campus version from other versions.
It was noted that any trademark changes would require collaboration between Axim and 2U, as 2U owns the edX brand. The community’s help would be needed to make the case that increased campus adoption benefits all parties, being good for education and for the brand overall. Legal counsel would be consulted to get up to speed on the request.
Development Status and Requirements
The campus version status included completed technical and architectural plans. The primary blocker was not a pull request or plugin, but rather that potential adopters want a complete project starting from branding, with clarity on who will support and maintain it over a long horizon (e.g., the next 10 years). This had been the top request from market interactions over the previous two years, and the business buy-in was being blocked by this uncertainty. Several community initiatives had been attempted previously but none had materialized.
Market Opportunity and Digital Sovereignty
Strong market signals were described. A recurring theme from convenings was a desire for digital public goods and digital public infrastructure not controlled by commercial organizations, especially US-based ones. A critical example cited was that the US government can turn off access to AI tools like Claude from one day to the next, indicating a lack of technological sovereignty—a concern extending to LMS and MOOC solutions. This created a moment for open source solutions to obtain funding from ministries and large funders like MasterCard.
It was also argued that AI is creating a resurgence in open source generally, as governments and companies work intensively on open source AI models, with some open source models claiming to approach parity with proprietary models from roughly three months prior, at an order of magnitude lower cost. The observation was made that open source had lost ground in US administrative systems over recent years, but a moment now exists—driven by Canvas’s weaknesses (including a data breach) and calls for digital sovereignty—to rebuild the brand as the clear choice for governments.
Lessons from Past Attempts
It was acknowledged that many attempts to build a campus solution had occurred previously with varying success, offering lessons and pitfalls to avoid. The platform was noted as being in a very different place from a plugin/API extensibility standpoint compared to earlier efforts. Reference was made to prior discussions about the platform serving as a “kernel” or “nugget” upon which use cases (or “bundles”) could be layered, with campus being one such example bundle. Optimism was expressed that closing the gaps was not insurmountable.
A conversation with the University of Bergen in Norway (currently using Canvas) was shared—they were interested in moving to an open source on-campus system. The honest assessment given was that the Open edX platform is “not yet there” for on-campus use, but the only way to know is to try it, suggesting they replicate one on-campus teacher’s course as a trial. The Studio/LMS separation was noted as excellent for MOOCs but requiring workarounds (custom X blocks, detached exams) for on-campus use where teachers upload files and set dates. The point was emphasized that changing an LMS faces significant teacher resistance—even upgrading to a better version encounters resistance—but the current moment of institutions considering moving from Canvas represents a good opportunity.
Path Forward and Leadership Need
A recommended sequence of three priorities emerged: first, a well-regarded key owner (potentially from Axim) needs to stand up as the visible face of the campus initiative (codenamed “Open edX Campus”); second, that owner with community help drives fundraising; third, a clear indication of the development gap and associated cost is needed.
It was noted that gap analysis specific to particular cases was already done (including studying the Spanish/edX-led case and MENA cases), while a complete survey of all interested parties remained outstanding but was not a blocker. Active bidding for national RFPs was already underway. A product management perspective examining specific customer needs and gaps was identified as important, with organizations like the German University of Digital Science cited as examples to consult.
The branding effort was framed as a long arc worth initiating now (a “forcing function”) but not a blocker for next steps. The critical near-term inputs identified were understanding the scope of changes needed on top of the core for a usable on-campus LMS and estimating the associated cost.
Open edX Security Brief [11:49:03 - 11:51:38]
Committee members were asked to comment asynchronously on the Open edX security brief. The focus was on ensuring the project stays ahead of AI’s evolution as both a threat and opportunity in the security space for open source applications. A specific request (limited to US organizations) involved seeking funding from the National Science Foundation, which has a program supporting sustaining open source projects with a security focus. Letters of support from US-based community members were needed for the grant application. MIT and WGU would be approached, and interest was expressed in having a commercial organization (2U) provide a letter of support, with up to five organizations possible (Clemson and UVA cited as potential additional candidates).
Contribution Visibility and Roadmap Funding Tool [11:51:38 - 12:04:34]
Concept Overview
A demo was shown of a community-built tool with two main purposes: giving organizations currently contributing to the open source project visible credit for their contributions in a new way, and making clear when roadmap features remain unfunded so that funders and organizations can accelerate work by bringing additional funding. An example scenario described a funder being shown a digital credentials feature that is planned but not yet funded, enabling them to sponsor its kickoff. Using dummy data, the demo illustrated how each roadmap item could display contribution details (e.g., development hours contributed by one organization, funded contributions from another, and the organization that built the feature).
Closing [12:04:34 - 12:04:52]
The meeting concluded a few minutes over time with appreciation for attendees’ participation. Notes would be summarized and distributed as usual, with a reminder to read and comment on the security draft.