Scope and preparation
This is a participant and facilitator workflow, checked against immutable Jitsi Meet 11248 source on 27 October 2026. It is an evergreen guide, not an announcement of a new polling feature. Private deployments, native apps and newer releases may show different layouts. No live poll, meeting or device test was performed. [S1][S2][S3]
Prepare a harmless practice question, such as which of three workshop subjects to discuss next. Tell attendees whether they should select one answer or all that apply. The reviewed answer form permits multiple choices; a written instruction does not enforce single-choice voting. [S2]
Do not use a casual meeting poll for a sensitive anonymous ballot without verifying its actual privacy properties. The reviewed results interface can reveal voter names, and Jitsi’s optional media end-to-end encryption does not cover polls. [S3][S5]
Create a draft, then send it
- Open the meeting’s Chat panel and select Polls when available. Some layouts also expose a Polls button. Availability depends on conference support and configuration. [S1][S4]
- Select Create a poll, enter the Poll Question and fill in at least two different, nonblank options. Use Add option for more answers or Remove option to remove extras. [S1]
- Select Save. The reviewed source saves a draft in the client’s poll state. This is the preparation step, not the conference send operation. [S1][S2]
- Review the saved card. Use Edit if the wording needs changing, then select Send. Its accessibility label is Send poll. This calls the conference poll API. [S2]
- Ask a second attendee in the same room to confirm receipt and read back the question before collecting responses. A local draft or outgoing request does not establish that everyone received the poll. [S2][S4]
If Save stays disabled, check for an empty question, fewer than two nonblank options or duplicate answer text. The reviewed form rejects these conditions. Use clearly distinct labels rather than trying to distinguish options using spaces alone. [S1]
The Close poll control on a saved draft removes that draft in the reviewed UI. Do not interpret it as proof that a poll already delivered to attendees has been withdrawn. The draft and the received poll have different states. [S2]
Vote, skip and change a response
Open Polls, locate the intended question, select the applicable answer checkboxes and choose Submit. The reviewed Submit button stays disabled until at least one answer is selected. Skip opens the results view without submitting a selection; it is not an abstention option added to the question. If you need to count abstentions explicitly, include an appropriate option when creating the poll. [S2][S3]
In the results view, Change vote returns a voter to their answer form. A person who skipped can use Vote. Confirm that the displayed response and result counts reflect the intended change rather than assuming a button click proves delivery. [S2][S3]
For keyboard problems while entering options or navigating controls, use the separate chat and poll keyboard guide. This page explains the normal workflow; it does not claim an open upstream accessibility proposal is a released fix.
Read the results correctly
The reviewed results show a count and percentage for each option, plus a Show details control that can list voter names. Hide details hides that list in your interface; it does not establish anonymous collection. Explain this before asking for personal or sensitive answers. [S3]
The normal calculation uses the number of voters for an option divided by the number of distinct people who voted across the options. Because one person can select several answers, percentages can add up to more than 100%. It is not a percentage of everyone currently in the meeting. For example, if two voters both choose two options, both options can display 100%. [S2][S3]
Treat these as informal meeting results, not a verified identity register or an audited election. The reviewed client clears its local poll state when the current conference changes. Do not rely on a recurring meeting link or leaving and rejoining as a durable result archive. If results must be retained, agree on an approved recording or note-taking process and verify it separately. [S4]
Missing polls, missing controls or failed delivery
| What you see | Check before changing anything |
|---|---|
| No Polls tab or button | The reviewed availability check requires conference polling support and a configuration that does not disable it with disablePolls. Ask the operator about the actual served client/configuration. [S4] |
| Polls appear, but no Create a poll | Creation has a separate check. Visitors cannot create polls in this source; dynamic branding can also require a create-polls JWT feature. Do not assume granting moderator status is the right correction. [S4] |
| Save is disabled | Check the question and at least two distinct nonblank answer options. [S1] |
| Only the author sees the new question | Check for a saved draft and the separate Send step, then verify delivery with another attendee. [S2] |
| The poll works in the main room but not a breakout | Compare the same harmless test in both rooms and record exact client/server versions. Do not assume a main-room poll and each breakout have identical state or support. See the breakout workflow. [S4][S6] |
| A participant gets no result after submitting | Check the correct room, the intended question, connectivity and whether a selection was made. Preserve a sanitized error and ask the operator to inspect poll support. An outgoing request is not server acknowledgement. [S2][S4] |
For self-hosted Docker and Debian/Ubuntu package deployments, first identify the installed version and inspect the configuration actually served to the browser. This article introduces no server-installation command, module patch or permission bypass. A working video call does not by itself prove that the deployment’s poll service is configured correctly. [S4]
Historical issue #16693 reported breakout polling failures in November 2025 and received an upstream correction. That history is a reason to check versions and room scope, not proof that your present failure has the same cause or that a newly opened proposal is a released fix. No current server reproduction was performed here. [S6]
Rehearsal and maintenance
Before using polls in an event, rehearse a draft, edit, send, receipt on another client, one answer, multiple answers, skip, change vote and result details. Include the actual attendee roles and native apps if those are required. Test a breakout separately if you intend to poll there. Use harmless answers and verify visibility before collecting anything sensitive. These checks are recommendations, not completed tests from this site.
Recheck this guide when the deployed release changes its draft/send flow, creation permissions, result privacy or poll-service behavior. The mobile guide and media E2EE guide cover different client and encryption scopes.
Primary sources
- [S1] Meet 11248 draft validation, web creation form and English labels.
- [S2] Same release: send, submit and skip actions and draft-versus-answer controls.
- [S3] Result calculation and vote changes, voter details interface and result-state reducer.
- [S4] Feature availability, creation permissions, Chat tabs and conference poll lifecycle.
- [S5] Official Jitsi security scope, checked 27 October 2026: media E2EE does not cover polls.
- [S6] Historical breakout poll report #16693, 21 November 2025, and merged correction #16702, 25 November 2025. Reviewed 27 October 2026; not a claim of a newly released fix or a completed meeting test.