Top 7 Open Source & Self Hosted Meeting Software Solutions for 2026

Somewhere around 2020, “open-source video conferencing” meant one thing: Jitsi, running on a spare server, occasionally falling over when more than a dozen people joined a call. Six years later, in 2026, that description does not hold up anymore. The category has split into genuinely different species of software, built for different jobs, and lumping them together in one comparison does a disservice to anyone trying to decide this quarter.

Some of these tools are meant to be installed once and clicked, no accounts, no infrastructure planning. Others are meant to be operated, patched, scaled and monitored by someone with a Kubernetes background. A few are not really “apps” at all, they are building blocks that companies use to construct their own meeting product, the way Ollasync and plenty of others have done. And a couple of them are not primarily about meetings, they are about encrypted messaging that happens to include calling as a feature.

This matters because the reason teams go looking for open source or self-hosted meeting software in the first place is rarely just “we want a free Zoom.” It is usually one of a handful of specific pressures: a compliance requirement that data cannot leave a particular jurisdiction, a security team that does not trust a third party server to see unencrypted video, a school or training company that needs breakout rooms and a shared whiteboard more than it needs enterprise polish, or a developer team that wants to build a meeting experience into their own product rather than embed someone else’s iframe.

So instead of ranking seven tools against one imaginary “best overall” score, this guide walks through what each one is built for, where it genuinely shines, where it will frustrate you, and who should pick it. We have used, deployed or closely evaluated all seven of these platforms, and we are not going to pretend any of them is perfect, because none of them are. Including, for what it is worth, our own product, which we will mention honestly near the end rather than pretend this is a neutral listicle when it clearly has skin in the game.

How we evaluated these tools

Before getting into the seven, it is worth being explicit about the criteria, because “best” without criteria is meaningless marketing language.

We looked at five things for each platform. First, setup and operational burden: how much server administration knowledge does running this thing in production require, and does it stay manageable as usage grows. Second, the meeting experience itself: audio and video quality under real network conditions, screen sharing reliability, and whether joining a call requires installing software or just clicking a link. Third, the feature set is relative to what it is aimed at, because a whiteboard and breakout rooms matter enormously if you run a virtual classroom and barely matter if you run investor calls. Fourth, the security and encryption model, specifically what a server operator can and cannot see, since “self-hosted” and “encrypted” are two entirely separate claims that get conflated constantly. And fifth, community health and maintenance velocity, because open-source software that has not shipped a meaningful release in two years is a liability no matter how good the architecture once was.

With that framing out of the way, here are the seven.

1. Jitsi Meet

Jitsi Meet is the tool most people picture when they hear “open-source video calling,” and for good reason. It has been around since 2003 in various forms, it runs entirely in the browser with no download required, and a participant can join a meeting by clicking a link the same way they would with any commercial tool. That single design decision, browser first with zero install friction, is why Jitsi remains the default starting point for teams testing the self-hosted waters.

Under the hood, Jitsi uses standard WebRTC for real time media. In its default configuration, calls under eight to ten participants use a full mesh topology, where every device sends its stream directly to every other device. That keeps latency low and avoids a central bottleneck, but it does not scale, because bandwidth and CPU load on each participant’s device grows with every additional person on the call. For anything larger, organizations deploy Jitsi Videobridge, a selective forwarding unit that receives each stream once and routes copies out to participants, which is what lets Jitsi deployments handle meetings running into the dozens of participants without every laptop fan spinning up.

The strengths here are real. It is genuinely free, genuinely open (GPL licensed, developed largely by 8x8 with an active community around it), and it is one of the easier platforms to get a basic instance running on a single server in the afternoon. Screen sharing, a text chat panel, virtual backgrounds, and basic moderator controls are all included out of the box. For a small team or a privacy conscious individual who just wants an alternative to Zoom for internal calls, it remains a very sensible first stop.

Where it gets harder is exactly where most “getting started” tutorials stop talking. Meeting scheduling, persistent rooms, user accounts and recording are not native concepts in a bare Jitsi deployment, because there is no coordinating backend by default, every room is just an ephemeral name. You can bolt on Jibri for recording and streaming, but Jibri is notoriously hungry and finicky to run reliably at scale, essentially requiring a dedicated headless browser instance per concurrent recording. Firewalls and restrictive corporate NATs can also cause connectivity issues that are genuinely difficult to diagnose without TURN server expertise. And because there is no default persistent identity layer, building anything resembling an enterprise deployment with SSO, room ownership and audit logs mean significant additional engineering on top of the open-source core.

Jitsi is best understood as excellent infrastructure with a thin, friendly front door. If your needs stay simple, that is a feature. If your needs grow into scheduling, compliance grade recording retention, and account-based access control, you will find yourself building most of that layer yourself or paying a hosted provider to have already built it for you.

2. BigBlueButton

If Jitsi is the general-purpose open-source meeting tool, BigBlueButton is the specialist, and it makes no secret of it. Everything about BigBlueButton is oriented around one job: running a live class or training session, not a generic business meeting. It has been developed since 2007 with that focus, and by 2026 it is difficult to find another open-source platform with a comparable depth of education specific tooling.

The feature list reflects the priority immediately. A real multiuser whiteboard where the presenter and, if permitted, students can annotate slides together in real time. Native breakout rooms that a moderator can create, populate and then pull everyone back with a single click, something Jitsi and most generic tools still require plugins or third-party services to replicate properly. Polling built directly into the presentation flow. Presentation upload that converts slides into a shared, annotatable canvas rather than a flat screen share. Shared notes that the whole class can edit together during the session. And a recording pipeline that captures not just video and audio but the whiteboard state and chat, stitched together into a coherent replay rather than a flat screen recording.

It integrates natively with the learning management systems that schools already run, particularly Moodle and Canvas, which is a substantial reason universities and training organizations gravitate toward it over a generic conferencing tool. Under the FreeSWITCH backed audio layer, voice tends to stay stable even when video streams are struggling on a poor connection, which matters a great deal in classroom settings where you would rather lose video than lose the teacher’s voice.

The tradeoffs are the ones you would expect from a mature, feature dense platform built primarily for one use case. Installation and ongoing server maintenance have a real reputation for complexity, users report needing a properly sized Ubuntu server with adequate RAM (four gigabytes tends to be treated as an informal minimum before video performance suffers) and the deployment process, while documented, has more moving parts than a simple Jitsi instance. Native client support outside the browser is thinner than some competitors, historically weaker for desktop Mac users and mobile compared to browser-based joining, though this has improved over recent releases including the 3.0 line that started rolling out to institutions in 2026. And because BigBlueButton is so purpose built for education, it can feel like overkill, or simply the wrong shape of tool, for a company that just wants sales calls and standups.

If you run a school, a bootcamp, a corporate training function, or any organization where “class” is a more accurate word than “meeting,” BigBlueButton in 2026 is still the most complete open-source answer available, and it is not particularly close.

3. Nextcloud Talk

Nextcloud Talk, sometimes still referred to by its underlying app name Spreed, takes a different angle entirely: it is not trying to be the best standalone meeting tool, it is trying to be the meeting layer of a broader self-hosted workspace. If your organization already runs Nextcloud for file storage, sharing and calendars, talk sits inside that same instance, using the same accounts, the same permission model and the same server you have already invested in securing and maintaining.

That integration is the entire pitch, and it is a highly compelling one for a specific audience. A calendar event in Nextcloud can spin up a Talk room automatically. A shared folder can have an associated chat and call thread attached to it. Files discussed in a meeting can be pulled up and co-edited without leaving the same browser tab. For organizations that have already committed to Nextcloud as their private cloud, and there are a great many of them across European public sector, healthcare and legal environments where data residency requirements pushed them there in the first place, adding Talk means one less vendor, one less identity system, and one less place data has to move between.

The call quality and feature set for straightforward meetings is genuinely solid: browser-based joining, screen sharing, basic call recording, and participant controls that cover what most day-to-day meetings need. Nextcloud has also invested in a High-Performance Backend option, which moves media routing off the standard signaling server and onto a dedicated component, meaningfully improving how many concurrent participants a Talk deployment can handle before performance degrades, though setting that piece up correctly is itself a nontrivial task.

The honest limitation is that Talk’s meeting features, on their own, are narrower than a dedicated conferencing platform. There is no equivalent to BigBlueButton’s classroom tooling, breakout rooms are more basic, and if you strip away the Nextcloud ecosystem around it and just ask “is this the best video calling experience available,” the answer is generally no. It is not trying to be. Talk earns its place on this list not because it is the most powerful meeting tool here, but because for the very large number of organizations already running Nextcloud for everything else, it is the path of least resistance to bringing meetings inside the same trust boundary as their files, without adding a second system to secure and audit.

4. Matrix and Element Call

Matrix is not, strictly speaking, a meeting product. It is an open, decentralized protocol for real time communication, originally built for messaging and later extended to cover voice and video calling through a component called Element Call, built on the group calling capabilities Matrix gained through its integration with LiveKit style SFU technology and, more foundationally, through the Matrix specific MLS based group encryption work the ecosystem has been maturing for several years.

The reason Matrix belongs on a 2026 evaluation of meeting software, rather than being filed purely under messaging, is that federation and end to end encryption are treated as first class design goals rather than bolted on afterthoughts. Because Matrix is a federated protocol, similar in spirit to email, different organizations can run their own homeservers and still communicate with each other without either party having to trust the other’s infrastructure directly. A call between someone on a government agency’s self-hosted Matrix server and someone on an NGO’s separate deployment can happen without either institution’s data ever passing through a shared third party. That is a genuinely different security posture from every centralized platform on this list, Jitsi, BigBlueButton and Nextcloud Talk included, all of which assume a single organization operates a single central server that every participant trusts.

The encryption story is the other half of the pitch. Modern Matrix rooms use group encryption schemes descended from the same cryptographic ratchet family that underpins the IETF’s Messaging Layer Security standard, meaning that even the homeserver operator, running their own infrastructure, cannot read message content or, in encrypted calls, the media itself. That is a meaningfully stronger guarantee than “encrypted in transit to our server,” which describes most of the other tools here.

What you give up for that is polish and simplicity, at least as of 2026. Element Call, while genuinely functional and improving steadily, still asks more of participants than clicking a link and joining, particularly around identity verification and cross device key management, concepts most non-technical meeting attendees have never had to think about with Zoom or Google Meet. Running your own homeserver well, with federation working correctly and reasonably fast media relay, is a more involved operational undertaking than standing up a single Jitsi box. And because the ecosystem is protocol first rather than product first, the experience can feel less unified than tools-built end to end by one team with one UX vision.

Matrix and Element Call are the right choice specifically when the encryption and federation guarantees are the actual requirement, not a nice to have. Government communication between agencies, journalist and source coordination, cross organization collaboration where no party wants to be the trusted central operator, these are the situations Matrix was built for, and in 2026 it remains close to unmatched at solving them.

5. LiveKit

LiveKit is the odd one out on this list, and it is worth being upfront about why it is here at all: it is not really a meeting application. It is open source, self-hostable, high-performance WebRTC infrastructure, a selective forwarding unit and accompanying SDKs that developers use to build real time audio and video features into their own products. If Jitsi and BigBlueButton are finished houses, LiveKit is a very good set of framing, plumbing and electrical work that a construction team uses to build whatever house they need.

That distinction matters enormously for those who should consider it. If you are a company that wants to add live video into an existing app, a telehealth platform, a marketplace with buyer seller video calls, a social product with group audio rooms, LiveKit is a serious, production grade foundation. It handles genuinely hard parts of real time media at scale: adaptive bitrate streaming, simulcast, server-side recording and compositing, egress to storage or live streaming platforms, and horizontal scaling across a cluster of media servers. It has SDKs across web, iOS, Android, Flutter and several backend languages, and a track record of being used by companies that need to build customs, branded real time experiences rather than adopt someone else’s meeting UI.

If, on the other hand, you are a team that just wants to hold internal meetings tomorrow, LiveKit is close to the wrong answer entirely. There is no box meeting room UI, no scheduling interface, no chat panel, no surface layer a non-technical team expects to open and use. Every one of those things must be built on top of the SFU using LiveKit’s client SDKs, which means the actual “meeting software” only exists once a development team has built it. That upfront engineering investment is precisely the tradeoff: you get complete control over the product experience and the ability to embed it exactly where you need it, in exchange for accepting that you are signing up to build and maintain an application, not install one.

We include LiveKit specifically because a meaningful share of people searching for open-source meeting software in 2026 are not looking for an app to hand to end users at all, they are technical teams evaluating what to build their own product on. If that is your situation, comparing LiveKit with finished products like Jitsi or BigBlueButton is comparing the wrong things, and it is worth recognizing that distinction before six months of engineering time goes into the wrong assumption.

6. OpenMeetings

OpenMeetings is the veteran of this list, an Apache Software Foundation project with roots going back well over a decade, and it occupies a slightly unusual space in 2026: still actively used, still maintained, but clearly built in a different design era than newer WebRTC native tools, and it shows in specific, sometimes charming, sometimes frustrating ways.

Its core strength has always been an integrated set of collaboration features bundled around browser-based video conferencing: real time chat, a shared whiteboard, document and presentation sharing, and configurable meeting rooms with granular participant permission settings, all managed through an admin interface that predates a lot of the “modern SaaS dashboard” conventions younger tools default to, for better or worse. It supports recording, and its room-based model with defined moderator and presenter roles maps well onto structured training sessions, webinars and formal meetings where you want tight control over who can speak, share their screen or use the whiteboard at any given moment.

For organizations that already run other Apache ecosystem tooling, or that value a project with a long, stable governance history over one built in the last few years, OpenMeetings offers a level of institutional maturity that is reassuring in procurement conversations, particularly in public sector and enterprise environments accustomed to evaluating software by longevity as much as by feature velocity.

The honest downside is that OpenMeetings can feel dated next to Jitsi or BigBlueButton in terms of interface polish and the sheer smoothness of the join experience, and development velocity, while ongoing, is slower than the more actively venture or foundation backed alternatives on this list. It is a tool that rewards an organization willing to configure it carefully and is less forgiving of a “just click deploy and hope” approach than some competitors. We would not recommend it as a first choice for a team starting from zero in 2026, but for an organization already invested in it, or one specifically prioritizing role based meeting control and long-term project stability over cutting edge features, it remains a legitimate, working option rather than a legacy dead end.

7. Jami

Jami is the purist on this list, and it deserves inclusion precisely because it approaches the entire problem from the opposite direction of everything else here. Where Jitsi, BigBlueButton, Nextcloud Talk and OpenMeetings all assume a central server that participants connect through, Jami is built as a fully distributed, peer to peer communication platform with no mandatory central server at all. It is a GNU project, licensed under the GPL, developed primarily by Savoir faire Linux with support from the Free Software Foundation, and its entire design philosophy is oriented around one goal: nobody, including the developers, including any server operator, sits in the middle of your conversation, because structurally there is no middle to sit in.

Calls, whether one to one or in a group, are negotiated and transmitted directly between participants’ devices wherever network conditions allow, using a distributed hash table for peer discovery rather than a central directory service. That architecture gives Jami a privacy story that is difficult for any centrally hosted tool, open source or not, to fully match, since there is no server log of who called whom, no central point that could be compelled to hand over metadata, and no infrastructure cost or operational burden of running conferencing servers at all, because there effectively are none to run.

The tradeoffs are exactly what you would expect from that architectural choice. Peer to peer topology works elegantly for one-to-one calls and reasonably well for small groups, but it does not scale to the dozens or hundreds of participants that a selective forwarding unit-based tool like Jitsi or BigBlueButton can handle, because without a central relay, adding participants adds direct connections and bandwidth load in a way that grows quickly. Features that assume a persistent server, like scheduled meeting rooms with a stable link participants join later, recording infrastructure, or large webinar style broadcasts, are either absent or fundamentally awkward to implement in a serverless model. And connectivity through restrictive corporate firewalls and complex NAT setups, always a challenge for peer-to-peer traffic, can be more inconsistent than with a centrally routed alternative.

Jami is the right tool for a specific, deliberate use case: individuals, journalists, activists or small teams who genuinely need the strongest achievable privacy guarantee for smaller conversations and are willing to trade away scale, scheduling convenience and polish to get it. For an organization’s weekly all hands with fifty participants and a recording archive, it is the wrong tool. For two people who need to be certain no server anywhere holds a record of their conversation, it is very close to the best open-source option available in 2026.

Comparing them without pretending there is one winner

Laid out side by side, a pattern becomes clear that a single ranked list would obscure. Jitsi Meet and Nextcloud Talk sit closest together as pragmatic, centrally hosted, browser first tools aimed at general purpose meetings, with Talk differentiated mainly by how deeply it integrates into an existing Nextcloud deployment rather than by raw conferencing power. BigBlueButton and OpenMeetings both aim squarely at structured, education and training oriented sessions, with BigBlueButton clearly ahead on modern polish and integration breadth, and OpenMeetings offering an older but stable alternative for teams already committed to it. Matrix with Element Call and Jami both prioritize a fundamentally stronger privacy and security model than the other five, but arrive there through opposite architectures, federation with strong end to end encryption for Matrix, versus no central server at all for Jami, and both ask more of their users in exchange. And LiveKit stands entirely apart as infrastructure rather than an application, the right answer is only for teams planning to build, not adopt.

The practical implication is that “which of these seven is best” is close to an unanswerable question as posed. The better question is which two or three of these categories match your constraints, and then which specific tool within that category fits your existing infrastructure and technical capacity.

Who should pick what

• Small teams & quick setups: If you are a small team or an individual who wants a private, self-hosted alternative to commercial video call tools for everyday internal meetings, and you are comfortable running one Linux server, start with Jitsi Meet. Accept upfront that scheduling, recording and account management will need additional work as you grow, and plan for that rather than being surprised by it later.
• Classrooms & online training: If you run a school, training company, bootcamp or corporate learning function where breakout rooms, a real whiteboard and LMS integration genuinely matter to how you teach, BigBlueButton is very likely your best available open-source option in 2026, and the deployment complexity is a worthwhile tradeoff for what it delivers.
• Existing Nextcloud workspaces: If your organization already operates Nextcloud for files and calendars, add Nextcloud Talk rather than introducing an entirely separate conferencing system, security boundary and login for meetings, unless your meeting outgrow what it offers.
• Federated & high-trust communication: If cross organizational communication with no shared trusted operator, or the strongest possible encryption guarantee for group calls, is a hard requirement rather than a preference, invest the extra operational effort in Matrix and Element Call. This is common in government, journalism, legal and cross border NGO contexts where trusting any single party’s server is itself the problem being solved (see our compliance guide for regulatory standards).
• Developers building video features: If you are a development team building real time video into your own product rather than looking to adopt a finished meeting app for internal use, evaluate LiveKit as infrastructure, not as a competitor to the other six, and budget real engineering time to build the actual user facing experience on top of it.
• Legacy enterprise & Apache stacks: If your organization has a long standing OpenMeetings deployment and it is doing its job, there is little urgency to migrate purely for novelty, though any fresh 2026 deployment should weigh it honestly against BigBlueButton’s more active development and broader feature set.
• Maximal privacy & zero servers: And if the requirement is genuinely maximal privacy for a small number of participants, with no tolerance for any server anywhere holding metadata about who spoke with whom, Jami remains the most architecturally serious open-source answer, provided your group size and need for persistent, scheduled meeting infrastructure stay modest.

The part of self-hosting nobody puts in the pitch deck

It is worth being directed about something the marketing pages for open-source meeting tools tend to gloss over: choosing self-hosted software does not automatically make you secure; it makes you responsible. Every one of the seven tools above needs someone who owns patching, capacity planning, TURN and STUN server configuration for reliable connectivity behind restrictive networks, backup strategy for recordings and metadata, and incident response if something does go wrong. A vulnerability disclosed against Jitsi Videobridge, BigBlueButton’s FreeSWITCH layer, or a Matrix homeserver implementation does not patch itself because the software happens to be open source, and an unpatched self-hosted server can end up considerably more exposed than a well-maintained commercial platform with a dedicated security team behind it.

This is not an argument against self-hosting; it is an argument for going in with clear eyes about the operational commitment attached to the security and control benefits. Organizations that treat “we self-host our meetings” as the finish line, rather than the starting point of an ongoing operational responsibility, are the ones who end up in the security incident postmortems a year later.

Where this leaves teams evaluating their options in 2026

We built Ollasync because a specific gap kept showing up in conversations with trainers, tutors and small teams evaluating exactly this landscape: they wanted the simplicity of a browser based, no install meeting tool, but they also wanted real end to end encryption available when a session genuinely called for it, live translation across languages for classes with international participants, and AI generated class notes that did not require choosing between convenience and control, deliberately excluding AI processing entirely from encrypted sessions rather than quietly compromising the encryption promise to make note taking work. That is a narrower, more opinionated bet than any of the seven general purpose platforms above, and it will not be the right fit for every situation this article covers, a fully peer to peer requirement is still better served by Jami, and a federated cross organization encryption requirement is still better served by Matrix.

But if you read through all seven of these and found yourself thinking “I want the ease of Jitsi, the education focus of BigBlueButton, and a real answer on encryption without needing a systems administrator,” that particular combination is exactly the space we are building in, and it is worth a look alongside the open-source options above rather than instead of them.

Frequently asked questions

Is open-source meeting software more secure than commercial platforms like Zoom or Google Meet?

Not automatically. Open source means the code is auditable and you control where it runs, which are genuine security advantages, but the actual security of a deployment depends on whether it is patched, configured correctly, and operated by someone who understands the threat model. A poorly maintained self-hosted server can be less secure than a well-run commercial platform. The advantage is potential and control, not a guarantee.

Can these tools handle large webinars with hundreds of participants?

BigBlueButton and Jitsi with a properly scaled Videobridge deployment can handle meetings well into dozens and, with sufficient infrastructure investment, hundreds of participants. LiveKit, as raw infrastructure, can scale further still because that is precisely the problem it is built to solve, but only once a team has built the application layer on top of it. Jami and, to a lesser extent, Matrix are not well suited to large scale webinars given their architecture.

Do I need to be a systems administrator to run any of these?

Jitsi Meet has the lowest barrier to a basic working deployment. BigBlueButton, Nextcloud Talk with the High-Performance Backend, Matrix homeservers and OpenMeetings all require meaningfully more server administration comfort, particularly as usage scales beyond a handful of concurrent meetings. LiveKit assumes a development team, not just a system administrator.

Which of these offers true end to end encryption, not just encryption in transit to the server?

Matrix and Element Call are built around end-to-end encryption as a core design principle, using group encryption schemes that keep even the homeserver operator from reading content. Jami’s peer to peer architecture means there is no central server to intercept traffic at all for most calls. Jitsi, BigBlueButton, Nextcloud Talk and OpenMeetings in their standard configurations encrypt media in transit to the server but the server itself is generally positioned to access unencrypted streams, which is an important distinction for anyone whose requirement is specifically end-to-end encryption rather than transport encryption.

Is it worth self-hosting at all in 2026, given how mature commercial platforms have become?

For teams with specific compliance, data residency or trust requirements, yes, self-hosting remains the only way to guarantee data never leaves infrastructure you control. For teams without those specific pressures, the operational overhead of running and securing your own meeting infrastructure is a real cost that should be weighed honestly against the convenience of a managed platform, commercial or otherwise, that has already solved the scaling and reliability problems for you.

Wrapping up

There is no single winner on this list, only seven tools built for seven different jobs. Jitsi Meet is the fastest way to get a private meeting tool running with minimal setup. BigBlueButton is the strongest choice for classrooms and training, with breakout rooms and a whiteboard built in. Nextcloud Talk makes the most sense if you already run Nextcloud and want meetings inside the same trusted workspace. Matrix with Element Call and Jami are the two real answers if end to end encryption or peer to peer privacy is a hard requirement, not a nice to have. LiveKit is infrastructure, not an app, and belongs only in a developer team’s toolkit. OpenMeetings remains a stable, if dated, option for organizations already invested in them.

Pick based on what you are solving, ease of setup, classroom features, encryption strength or scale, not on which name sounds most familiar. And remember that self-hosting shifts responsibility to you: patching, capacity and security become your job the moment you deploy any of these.

If what you want is something in between, simple enough for a non-technical trainer to use, but with real encryption, sovereign deal rooms, and AI notes when a session calls for it, that is the specific gap Ollasync is built to fill. To compare self-hosted security with commercial cloud providers, see our transparent pricing and deployment resources.

Related reading

• Deployment — How to Deploy an On-Premises Video Conferencing Server in 30 Minutes: A step-by-step guide to standing up production self-hosted video infrastructure.
• Security — How to Run Air-Gapped Video Conferencing for Maximum Internal Security: How high-security teams run meetings completely isolated from public networks.
• Branding — How to Host Custom Domain Video Meetings Under Your Own Brand URL: Setting up dedicated single-tenant and branded meeting infrastructure.
• Compliance — Video Conferencing Compliance for Finance and Legal Teams: Risks, Requirements, and Solutions: Meeting regulatory obligations under SEC, FINRA, MiFID II and GDPR.