Opensource.com
The Fedora Design Team proved that an open source toolchain can handle a full virtual conference when it hosted the first Creative Freedom Summit in January 2023.
This article covers the conference’s background, why we prioritized open source tools, the exact stack we used, and what we’d change for the next summit in 2024.
What is Creative Freedom Summit?
The Creative Freedom Summit started as an idea from Marie Nordin after she reviewed talk submissions for Flock, the annual Fedora users and contributors conference. The August 2022 Flock received far more design and creativity-related submissions than it could accept, and Marie saw an opportunity for a separate conference dedicated to creatives who use open source tools.
Marie brought the idea to the Fedora Design Team in fall 2022, and planning began for the event held January 17-19, 2023. Since this was our first time running a conference of this kind, we started with invited speakers drawn from the Flock submissions and our own network of open source creatives. Nearly every speaker we asked accepted, so we had no room for an open call for papers. We’ll need to revisit that for next year, so we don’t yet have an open source CFP tool to recommend.
Why use open source for an open source conference?
Since the early pandemic lockdowns, Fedora’s Flock conference has run virtually on Hopin, a proprietary platform that integrates well with open source tools. Fedora adopted it years ago, and it provides a polished conference experience with a sponsor expo hall, tracks, hallway chat, and moderation tools. Choosing Hopin for the Creative Freedom Summit was technically possible since Fedora’s existing Hopin setup was available to us.
As a ~20-year open source contributor, I can tell you this decision is never easy. Hosting an open source-focused conference on a proprietary platform feels misaligned. But as community events grow in scale and complexity, building an integrated open source conference system becomes harder.
There is no single right answer. The choice depends on weighing several factors:
- Budget
- People power
- Infrastructure
- Technical capability
- Event complexity, formality, and culture
We had no budget for this event. We did have volunteers willing to contribute time, the Fedora Matrix Server as supported infrastructure, and access to a hosted WordPress system for the website. Teammate Madeline Peck and I had experience running the live, weekly Fedora Design Team video calls via PeerTube. We also wanted the event to be low-key, single-track, and informal, which lowered our tolerance for complex tools. Most importantly, we had strong motivation to try an open source stack.
Understanding these considerations may help you make your own decision for your event.
An open source conference stack
Here’s how the conference tech stack was assembled.
Live components
- Livestream: We streamed the stage and social events to a PeerTube channel. Attendees watched live from the channel, which includes privacy-minded analytics for viewer counts and post-event views.
- Live stage + social event room: One Jitsi room served as the live stage, with camera access restricted to speakers and hosts. A second Jitsi room hosted social events where any attendee could turn on their camera.
- Backstage: A private Matrix channel coordinated speakers, hosts, and volunteers during the event.
- Announcements and Q&A: A shared Etherpad (later moved to Hackmd.io) managed the daily schedule and collected audience questions.
- Integrated conference experience: Using Matrix’s Element client, we embedded the livestream video and an Etherpad into the main public Matrix room. Attendance in that channel served as our conference attendance metric, and live chat questions came from both the chat and the embedded Q&A pad.
- Conference website: Ryan Gorley designed a WordPress site with basic event info, dates and times, and the schedule.
Post-event components
- Post-event survey: We used the open source LimeSurvey system to gather attendee feedback, and some of that data appears in this article.
- Post-event video editing and captioning: We didn’t have live captioning, but I typed live notes into the chat during talks, which attendees appreciated. After the event, we used Kdenlive (one of the tools featured in talks) to edit videos and generate captions.
- Event recordings: PeerTube automatically posts livestream recordings to channels, making near-instant replays available for attendees who missed sessions.
Details on each component follow.
Livestream with PeerTube
(Máirín Duffy, CC BY-SA 4.0)
We used the LinuxRocks PeerTube platform, generously hosted by LinuxRocks.online, for the summit’s livestream. PeerTube is a free, open source, decentralized video platform that’s part of the Fediverse.
A standout PeerTube feature—one I haven’t seen elsewhere—is the near-instant replay recording that posts to your channel after a livestream ends. Attendees in our chat cited this as a major advantage. Someone who missed a session could watch it within minutes of the talk’s conclusion, with no manual uploading or coordination required from the organizing team.
Here’s how livestreaming with PeerTube works: create a new livestream on your channel, and you receive a livestreaming URL plus a key to authorize streaming. This URL and key can be reused, and we configured it so recordings post to the channel automatically when a livestream ends. You then copy and paste this into Jitsi when you start the livestream. This reuse saved organizers significant overhead—we didn’t need a new URL and key for each talk. Instead, we shared one URL and key in a common document among the organizing team (we each had different hosting shifts), and anyone with access could start the stream.
How to generate the livestream URL + key in PeerTube
The following steps cover generating the livestream URL and key in PeerTube.
1. Create stream video on PeerTube
Log into PeerTube, and click the Publish button in the upper right corner:
(Máirín Duffy, CC BY-SA 4.0)
2. Set options
Click on the Go live tab (fourth from the left) and set these options:
- Channel: (The channel name where you want the livestream to publish)
- Privacy: Public
- Radio buttons: Normal live
Then, select Go Live. (You’re not actually going live yet—there’s more data to fill in.)
(Máirín Duffy, CC BY-SA 4.0)
3. Basic info (don’t click update yet)
First, fill out the Basic Info tab, then move to the Advanced Settings tab. Fill in the livestream name, description, tags, categories, license, and other details. Be sure to check the publish-after-transcoding box.
This setting ensures the recording automatically posts to your channel when the livestream ends.
4. Advanced settings
You can upload a “standby” image that displays while viewers wait for the stream to start.
(Máirín Duffy, CC BY-SA 4.0)
Here’s the standby image we used for the Creative Freedom Summit:
(Máirín Duffy, CC BY-SA 4.0)
5. Start livestream on PeerTube
Select the Update button in the lower right corner. The stream will appear in a holding pattern until you start streaming from Jitsi:
(Máirín Duffy, CC BY-SA 4.0)
6. Copy/paste the livestream URL for Jitsi
This is the final step in PeerTube. Once the livestream is up, click the … icon under the video toward the right:
(Máirín Duffy, CC BY-SA 4.0)
Select Display live information. A dialog like this appears:
(Máirín Duffy, CC BY-SA 4.0)
Copy both the live RTMP URL and the livestream key, then combine them into one URL to paste into Jitsi.
Here are examples from my test run of the two text blocks:
- Live RTMP URL: rtmp://peertube.linuxrocks.online:1935/live
- Livestream key: 8b940f96-c46d-46aa-81a0-701de3c43c8f
Paste them into Jitsi combined with a / between them, like this:
rtmp://peertube.linuxrocks.online:1935/live/8b940f96-c46d-46aa-81a0-701de3c43c8f
Live stage + social event room: Jitsi
We used the free and open source hosted Jitsi Meet for our live stage. We created a meeting room with a custom URL at https://meet.jit.si and shared it only with speakers and organizers.
We enabled the lobby feature (available in meeting settings once you join the room) so speakers could enter a few minutes early without interrupting the previous presentation. Hosts admitted them when the prior session ended. A password is another option, but the lobby sufficed for us. One caveat from testing: moderator status didn’t persist. If a moderator left the room, they lost moderator status and settings like the lobby. I kept my Jitsi room open for the entire conference to avoid this. Your experience may differ.
Jitsi has a built-in livestreaming option where you can post a URL to a video service and stream your meeting there. We trusted this approach because it’s how we run the weekly Fedora Design Team meetings. For the summit, we connected the Jitsi Live Stage to a channel on the Linux Rocks PeerTube.
Jitsi also lets speakers share their screens to drive slides or live demos.
Livestreaming Jitsi to PeerTube
1. Join the meeting and click the … icon next to the red hangup button at the bottom.
(Máirín Duffy, CC BY-SA 4.0)
2. Choose Start live stream from the pop-up menu.
(Máirín Duffy, CC BY-SA 4.0)
3. Copy/paste the PeerTube URL + key text.
(Máirín Duffy, CC BY-SA 4.0)
4. Listen for your Jitsi Robot friend.
A voice will announce, “Live streaming is on,” within a few seconds. Once you hear it, you’re live.
5. Stop the livestream.
Stopping the Jitsi stream deactivates the PeerTube URL, so repeat these steps to start again.
Jitsi tips
Managing recordings by turning the Jitsi stream on and off
We learned that turning the Jitsi stream off between talks produces one raw recording file per talk on PeerTube. On the first day, we let it run continuously, which resulted in some videos containing multiple presentations. This made the instant replay function less useful—attendees had to seek within a video to find their talk or wait days for the edited version.
Preventing audio feedback
We hit an audio feedback issue live that never appeared during testing, and it was entirely my fault (apologies to everyone who attended). I was setting up the Jitsi/PeerTube links, monitoring streams, and helping host. Even though I knew to mute PeerTube browser tabs once live, I either had more tabs open than I realized or the livestream autostarted in my Element client, which I had open to monitor chat—and Element has no easy mute. You may hear some rushed speaker intros from me; I knew I had about 30 seconds before feedback started.
Simpler ways to avoid this:
- Don’t have the host/emcee also be the person setting up and monitoring streams and chat. (Not always possible with limited volunteers.)
- If possible, monitor streams on one computer and emcee from another. That gives you one mute button on the monitoring machine and simplifies hosting.
This is worth practicing well ahead of your event.
Backstage: Element
(Máirín Duffy, CC BY-SA 4.0)
We created an invite-only Backstage chat room about a week before the conference and invited all speakers. This served two purposes:
- Speakers were onboarded to Element/Matrix before the event, with time to resolve any signup issues (none arose).
- We established a live communication channel with all speakers in advance, making announcements straightforward.
During the event, the channel coordinated speaker transitions, flagged schedule delays, and handled one emergency reschedule when a speaker couldn’t make their slot.
We also set up a separate room for hosts, but it turned out to be unnecessary—we just used the backstage channel. Two channels were manageable; three felt like too many.
Announcements and Q&A: Etherpad/Hackmd.io
(Máirín Duffy, CC BY-SA 4.0)
We pinned a widget in our main Element channel with general event information, including the daily schedule and code of conduct. Each day’s section had a per-talk area where attendees could drop questions for Q&A, which the host read aloud to the speaker.
After the first day or two, some attendees reported the Etherpad widget failing to load. We switched to an embedded hackmd.io document pinned as a widget, which worked better. We never fully diagnosed the original loading issue, but posting the raw (non-embedded) link in the channel topic gave attendees a workaround.
Integrated and centralized conference experience
(Máirín Duffy, CC BY-SA 4.0)
Matrix via Fedora’s Element server was the central hub for attending the conference. Matrix chat rooms in Element support a widget system that embeds websites directly into the room, which was key to making our channel the single place to be.
We embedded the PeerTube livestream into the channel (upper left in the screenshot). After the event, we shared a playlist of unedited video replays through the same interface. Now that the editing project is complete, the channel hosts a playlist of the edited talks in order.
As mentioned earlier, we embedded a hackmd.io note in the upper right for the day’s schedule, announcements, and Q&A. I wanted to build a Matrix bot for Q&A but couldn’t get one running—it’s a candidate project for next year.
Conference conversations happened in the main chat below these widgets.
A few considerations when using Matrix/Element as your central conference hub:
- The best experience is on the Element desktop client or a desktop web browser. The Element mobile client does support widgets, though some attendees found the UI unintuitive, and other Matrix clients may not display widgets at all.
- Attendees can easily assemble their own experience. Users who didn’t use Element reported no issues joining the chat or opening the PeerTube livestream URL directly. We shared the livestream and hackmd URLs in the channel topic as an alternative entry point.
Website
(Máirín Duffy, CC BY-SA 4.0)
Ryan Gorley built the Creative Freedom Summit website on WordPress. It’s hosted by WPengine, a one-pager with the conference schedule embedded from sched.org.
Post-event
Post-event survey
We used the open source survey tool LimeSurvey, sent to attendees via the Element Chat channel and our PeerTube video channel within a week or two after the event. Organizers continue to meet regularly, and we develop survey questions in a shared hackmd.io document. Here’s what we learned for your own open source-powered conference:
- Most attendees learned about the event from Mastodon and Twitter (together, 70% of respondents).
- 33% attended via the Element desktop app, and 30% used the Element web app—roughly 63% used the integrated Matrix/Element experience. The rest watched directly on PeerTube or viewed replays afterward.
- 35% of attendees made connections with other creatives through the chat, confirming that chat is essential for networking.
Captioning
Attendees appreciated the live chat notes and expressed interest in live captioning for better accessibility. Our stack didn’t include live captioning, but open source options exist. One is Live Captions, covered by Seth Kenlon in an opensource.com article, Open source video captioning on Linux. While that tool is intended for local viewing, a conference host could run it and feed it into the Jitsi livestream via the open source broadcaster OBS so all viewers benefit.
Post-event, we found that Kdenlive, our open source video editor, can automatically generate and place subtitles. Basic instructions are in the Kdenlive manual. Fedora Design Team member Kyle Conway, who helped with post-event editing, created a comprehensive tutorial (including video) on automatically generating and adding subtitles in Kdenlive. It’s worth a look if you’re interested in this feature.
Video editing volunteer effort
After the event, we recruited volunteers from the conference Element channel to edit videos—adding title cards, intro/outro music, and general cleanup. Some automatic replays were split across two files or combined multiple talks into one file, requiring reassembly or cropping.
We organized the work using a GitLab epic, with an FAQ and call for help organized by skillset, plus issues attached to each video. Custom labels tracked each video’s status and the type of help needed. All videos are now edited, though some still need descriptions written. Many have auto-generated subtitles that haven’t been proofread for spelling or other errors common with auto-generated text.
(Máirín Duffy, CC BY-SA 4.0)
Since video files could be large, volunteers downloaded raw recordings from the main PeerTube channel, edited them locally, and uploaded the finished versions to a private PeerTube account. Admins with access to the main channel periodically moved videos from the private account to the main one. PeerTube doesn’t support multiple accounts on a single channel, so we shared the main password with a few trusted admins. It’s not ideal, but it limited password exposure while letting volunteers contribute without friction.
Ready to give it a try?
I hope this detailed look at how we ran the Creative Freedom Summit with an open source stack inspires you to try it for your event. Let us know how it goes, and reach out if you have questions or suggestions. Our channel is at https://matrix.to/#/#creativefreedom:fedora.im.
This article is adapted from Run an open source-powered virtual conference and is republished with permission.



Comments are closed.