Run Sheet: A 90-Minute Language Exchange With Live Captions on a Second Screen

Prepare the room before the first arrival

A language exchange is a social event with a teaching element. Participants need time to find one another, understand the format and recover from the occasional technical problem. Before hosting a virtual event in social VR, settle the room layout, voice arrangements, conversation pairings and caption access instructions. A second screen is useful because captions can remain visible without taking over the virtual environment. Participants using desktop clients may place the caption window beside the world, while headset users may need a companion device.

Prepare a short welcome message explaining the event's purpose, the languages being practiced and the expected conversation format. State that automated captions can omit words, misunderstand accents and struggle with overlapping voices. Participants should be able to ask the host for clarification without feeling that they have interrupted the event. If the platform requires a separate caption service, explain what audio it receives and who can access the resulting text.

• Set up a visible welcome point and a quiet help area.
• Prepare a participant list and rotation plan.
• Provide caption instructions before arrival.
• Identify a backup host who can take over timing.

The 90-minute event schedule

A predictable schedule gives newcomers confidence and prevents experienced speakers from dominating every conversation. The following plan assumes one lead host, one helper and a manageable group that can be divided into pairs or small groups. If attendance exceeds the available conversation spaces, create a waiting area and explain how the next rotation works.

• 00:00-00:10: Welcome, orientation and audio checks. The host explains the format, introduces the caption display and answers access questions. The automated caption helper can provide text, but a person handles individual assistance.
• 00:10-00:20: First-language introductions. Participants share their names, preferred languages and one simple conversation topic. The host watches for newcomers who have not found a partner.
• 00:20-00:35: First conversation round. Pairs practice the first language. Captions remain available on the second screen, while the host handles timing and requests to change partners.
• 00:35-00:40: Rotation break. Participants move to new partners. The host announces the next language and gives everyone time to settle.
• 00:40-00:55: Second conversation round. Participants switch practice languages. A volunteer can help visitors who are having trouble hearing, reading or moving between groups.
• 00:55-01:10: Small-group discussion. Groups of three or four discuss a shared prompt. The host reminds participants to speak one at a time so captions remain easier to follow.
• 01:10-01:20: Open conversation. Participants choose a topic and partner. The host stays available for corrections, accessibility requests and misunderstandings.
• 01:20-01:30: Closing and next steps. The host summarizes announcements, shares the next event time if known and explains whether any event materials will be retained.

Put captions on a second screen without making them a second event

The purpose of the second screen is to make spoken conversation easier to follow, not to create another interface that everyone must constantly watch. Use a readable font, strong contrast and a stable window position. Avoid displaying a long stream of historical captions when participants only need the current conversation. If the caption service offers speaker labels, explain that those labels may be inaccurate and should not be treated as proof of identity.

For example, a participant using a desktop computer might keep the virtual world on the main display and place the caption window on a second monitor. A headset user might use a phone or tablet. Ask participants to confirm that they can read the text before the first round. If a service introduces noticeable delay, tell the group to allow a pause after each speaker finishes. Do not ask participants to treat captions as an exact transcript or a substitute for direct clarification.

Keep the human host responsible for the room

Automation is most useful when the task is repetitive and the consequences of an error are limited. Captions can display speech, a reminder system can announce a rotation and a scheduled message can direct late arrivals to the welcome area. These functions do not make an automated helper responsible for the social experience. A human must still decide how to handle a participant who is distressed, a disagreement about a translation or a request to leave a conversation.

Assign the helper specific duties before opening the event. One person can manage the schedule while another watches arrivals and accessibility requests. If the group is small, the lead host can perform both roles, but should avoid moderating a sensitive dispute while also relying on an unattended automated announcement system. Have a simple manual fallback for every essential function.

Common problems and practical responses

Caption errors are especially disruptive when participants are learning unfamiliar vocabulary. Encourage speakers to repeat an important phrase slowly, then let the listener ask a question. Do not make a participant defend an accent or speaking style because a caption has rendered a sentence incorrectly. If the service stops working, continue with spoken conversation and offer written clarification through a human helper.

Overlapping conversations can also make shared caption feeds difficult to interpret. Separate groups physically where the world allows it, or use individual conversation areas with distinct audio arrangements. Check the platform's current voice and caption capabilities before promising that separate groups will receive separate caption streams. Product behavior can change with client versions and service settings.

• If captions lag, announce a short pause between speakers.
• If a caption is unclear, ask the speaker to repeat the phrase.
• If a participant cannot access captions, offer direct human assistance.
• If the caption service fails, continue with the prepared manual format.

Close the event with a useful debrief

A good closing takes less than ten minutes and gives the host enough information to improve the next gathering. Ask whether the rotations were long enough, whether the caption window was readable and whether participants understood the language changes. Collect feedback voluntarily. Avoid retaining detailed conversations or participant audio simply because a caption service made that information available.

After visitors leave, the host and helper can record operational notes: attendance range, rotation problems, caption interruptions, accessibility requests and changes needed for the next session. Keep these notes focused on the event rather than evaluating individual speakers. A short record of practical decisions is more useful than an unnecessary transcript.

• What time did the first participant arrive, and when did the event begin?
• Which rotation or transition caused the most confusion?
• Could participants access and read the captions comfortably?
• Which task needs a named human owner next time?
• What single adjustment should be made before the next event?

Questions readers ask