To add live captions to a YouTube stream with SyncWords, send your existing broadcast to SyncWords Live, configure transcription and a captioned RTMP output, then deliver that output to YouTube. YouTube Studio must be set to receive embedded captions, and the SyncWords service must be running for the handoff to work.
This is a guide to SyncWords’ documented YouTube workflow, not a hands-on test of a live stream or the service interface. It covers the roles of your existing streaming software, SyncWords, and YouTube; check the current documentation and your own account settings before relying on the configuration.
Check the YouTube and SyncWords workflow
The key is to treat this as a relay with two separate connections, not as a captions switch inside your encoder. Your existing streaming software sends video and audio to SyncWords. SyncWords processes the incoming audio for captions and sends a captioned RTMP stream onward to YouTube.
The main steps are to create or open a SyncWords Service, choose RTMP PUSH as its input, and use the generated input details in your current streaming software. Then configure transcription, set the RTMP PUSH output and its YouTube destination, and enable Embedded 608 captions. In YouTube Studio, enable Closed Captions and select Embedded 608/708 as the Captions Source. Finally, start the SyncWords Service so it can accept input and deliver output.
SyncWords’ YouTube workflow instructions describe this sequence. It matters that both ends agree on the caption format: SyncWords warns that if YouTube’s Closed Captions setting is not enabled and Embedded 608/708 is not selected as the source, the stream passed to YouTube will not contain captions through this route.
Your encoder still has a job, but it no longer sends directly to YouTube. When you switch its destination to SyncWords, YouTube receives the output configured there instead. Keep the YouTube Stream URL and Stream Key available for the output setup, and keep the SyncWords input details available for the encoder. They belong to different legs of the route and should not be interchanged.
If you are starting from an existing OBS loop, review the OBS stream key settings for an always-on loop before changing its destination. A stable source stream is useful groundwork, but it does not configure captions or YouTube’s caption source by itself.
Create or open a SyncWords Service
Sign in to SyncWords and create a new Service or open the existing Service you intend to use. A Service is the workflow that accepts the incoming stream and applies the configured processing and outputs. The YouTube directions also refer to starting the Service separately; saving settings is not the same as making it ready to receive the live input.
Before you edit anything, identify which broadcast will feed it and whether the stream is already going to another destination. Record the current encoder destination and settings so that you can restore them if needed. If you use a scheduled playlist or a relay managed by someone else, agree who will change the output target and who will start the SyncWords Service. Two people changing destinations independently can leave you unsure which system is receiving the broadcast.
Think through caption languages at this point. Same-language captions, translated subtitles, and dubbed audio are different output choices. SyncWords’ output documentation ties buffer guidance to these choices, so avoid enabling translation or dubbing merely because the option exists. Decide what viewers need, then configure only the processing and outputs for that use.
You can also plan whether the stream needs a single YouTube destination or more than one RTMP destination. SyncWords documents up to four simultaneous RTMP destinations, with destination details configured separately. That is a documented capability, not a requirement for this YouTube setup. A single destination is easier to check when you are establishing the caption path for the first time.
Choose RTMP PUSH under Input Media
In the Service, open Input Media and select RTMP PUSH. SyncWords provides input details for the encoder or streaming service to push to. Copy the generated server URL and stream key or equivalent connection fields carefully into the existing streaming software, following the labels shown in the SyncWords account. Do not substitute YouTube’s Stream URL or key here: those are for SyncWords’ output connection.
SyncWords’ YouTube guide specifies input parameters for its documented RTMP 608 workflow. It calls for an audio sample rate of 44.1 kHz, a frame rate of 29.97 or 30 frames per second, a minimum video bitrate of 3,500 Kbps, a keyframe interval of two seconds, and an aspect ratio of 16:9 or 4:3. Attribute these values to the guide; they are not independent test results or a promise that every encoder configuration will work. Check the official SyncWords instructions for the current details and compare them with the controls available in your encoder.
These values should be considered together with the existing stream. For example, changing the keyframe interval or frame rate in one place may alter what the receiving service gets; changing only the output does not necessarily correct an incompatible input. If your current stream differs from the documented parameters, decide whether to adjust the encoder or seek clarification from SyncWords before the event rather than discovering a mismatch during the broadcast.
Once the input target is set, verify the source software is configured to send the programme’s audio as well as video. Captions are generated from incoming audio, so a silent or wrong-audio input will not yield the spoken or sung words you expect. An encoder preview can help establish that the intended programme is being sent, but it is not proof that captions are reaching YouTube.
Send the stream from existing streaming software
Use the RTMP PUSH details from SyncWords as the destination in the software or service already producing the live broadcast. SyncWords names OBS, Wirecast, and StreamYard as examples in its guide. The task is a destination handoff: preserve your normal media source and encoding configuration where possible, and direct the outgoing stream to the SyncWords input rather than the YouTube endpoint.
For OBS, that means using the custom server and stream key fields for the SyncWords input, rather than leaving the YouTube destination active. Other software may label the fields differently, but the distinction remains the same: input credentials belong to the stream sender, while YouTube credentials belong to SyncWords’ RTMP output. Keep a private record of each set in a secure place; stream keys grant access to a broadcast destination.
Start the source stream only after checking the input target, audio, and video settings. Depending on the order and service state, you may need to start the SyncWords Service before it accepts the incoming stream. SyncWords explicitly says its Service must be started to accept incoming streams and deliver an output. The exact interface sequence may change, so use the current account controls and documentation rather than relying on an assumed button location.
Operators who run a repeating playlist should also confirm that the programme remains continuous at the point where it enters SyncWords. Guidance on keeping a YouTube live stream active between playlist videos can help you inspect gaps in the source schedule. In this captions workflow, the relevant question is whether the source keeps sending the expected audio and video into SyncWords, not just whether the playlist is still open on a computer.
Configure YouTube delivery in SyncWords
In the Service’s Output settings, choose RTMP PUSH and add a destination for YouTube. Use the Stream URL and Stream Key from YouTube Studio’s Stream Settings in the destination fields. Keep the destination credentials separate from the SyncWords input credentials; the first connection goes from your encoder to SyncWords, and the second goes from SyncWords to YouTube.
Under Captions & Subtitles, enable Embedded 608 for the output. SyncWords describes 608 captions as traditional closed captions embedded directly into the video signal. Configure transcription so the incoming audio produces the caption language you need. Merely selecting an output format does not itself create transcript text; transcription needs to be set up as part of the Service.
SyncWords also documents assigning languages to CC1 and, if needed, CC2 through CC4. Do not assume every player can read captions from the extra channels: SyncWords cautions that support varies. For a straightforward YouTube workflow, begin with the caption channel and language you need, then verify the destination behaviour before depending on additional channels. Its output guide also describes a passthrough choice for preserving CC1 captions already present in the source; that is different from generating new captions from audio.
YouTube’s settings are a matching part of the handoff. In YouTube Studio, open the live stream’s Stream Settings, obtain its Stream URL and Stream Key for the SyncWords destination, and enable Closed Captions with Embedded 608/708 selected as the Captions Source. YouTube’s Live Control Room help explains the live control workflow; consult the current settings in Studio because labels and available controls can change.
Buffering is a trade-off, not a universal quality setting. SyncWords’ RTMP output guide recommends a 60-second buffer when using translated audio, 30 seconds for translated subtitles only, and between zero and five seconds for low-latency same-language captions. These are the guide’s recommendations for different workflows, not an unconditional setting. More buffering can accommodate processing and translation choices but adds delay; a lower buffer better suits a same-language, low-latency goal.
| Caption workflow | SyncWords buffer guidance | Practical consideration |
|---|---|---|
| Same-language captions with low latency | 0–5 seconds | Favour a prompt caption path where translation is not needed. |
| Translated subtitles only | 30 seconds | Allow for the translation workflow and its added delay. |
| Translated audio | 60 seconds | Account for dubbed audio processing and a longer delay. |
SyncWords says its RTMP output can send to up to four destinations simultaneously. The destinations share the encoded stream and caption set, while connector type, destination credentials, and dubbing language may differ. If your task is YouTube only, adding extra destinations increases the number of endpoints and credentials to check without being necessary for the documented YouTube route.
Verify captions in the live workflow
Before a public programme or a long-running channel depends on the caption path, verify each handoff in order. First confirm the source sends the intended audio and video to SyncWords. Then confirm the Service is started and that its output is directed to the intended YouTube live event. In YouTube Studio, verify the Closed Captions control and Embedded 608/708 source selection. Finally, inspect the live viewing experience for caption display and timing.
Do not treat a visible video preview as evidence that captions are present. Video can arrive while the audio track is missing, transcription is not configured, or the embedded caption setting is off. Likewise, a transcript visible in a SyncWords control or preview would not alone demonstrate that YouTube is receiving the embedded caption signal. Check the viewer-facing result in the actual YouTube stream before announcing that captions are available.
Use a short private or unlisted rehearsal when feasible, with representative speech, music, and pauses. Listen for whether the audio is present and observe whether captions appear, disappear, or lag in a way that affects the programme. This is an operational check, not a guarantee of accuracy or synchronization. Names, devotional terms, local place names, and code-switching can be difficult for automated transcription; prepare a person to monitor the output if those details matter to viewers.
For a 24/7 stream, plan a check after any restart, source change, or caption configuration edit. A restart can affect the service handoff even if the schedule looks unchanged. If your source runs from a home PC, the recovery steps for a playlist stream during Indian power cuts are relevant to restoring the upstream stream, but caption delivery still needs its own check after service resumes.
Troubleshoot the service handoff
When captions are absent, isolate the route from left to right rather than changing several settings at once. Confirm the encoder is publishing to the SyncWords input details, the Service is started, transcription is configured, the RTMP PUSH output has the correct YouTube URL and key, Embedded 608 is enabled, and YouTube Studio has Closed Captions enabled with Embedded 608/708 selected. Each item resolves a different possible break in the chain.
If YouTube shows video but no captions, focus first on the paired caption settings: Embedded 608 in SyncWords and the YouTube caption source. Then check whether the input contains the intended audio and whether transcription is configured for it. SyncWords’ workflow documentation says the captions will not be present in the passed-through stream when YouTube’s Closed Captions option and the Embedded 608/708 source are not set as required.
If YouTube receives no stream, distinguish an input failure from an output failure. Check whether the source is sending to the SyncWords input, whether the Service is active, and whether the output destination uses the YouTube credentials for the correct live event. An expired or mismatched key, a stopped source, or a Service that has not been started can break different parts of the route. Avoid pasting a key into the wrong endpoint while trying to recover.
If captions are delayed, first identify whether the workflow uses translated subtitles or dubbed audio, then review the relevant buffer choice in SyncWords’ output configuration. The documented buffer recommendations differ by workflow. Reducing buffering may reduce delay, but it does not resolve transcription, translation, or source-audio problems; observe the result after changing a single setting.
If captions appear inconsistently across devices, consider whether you assigned them to CC2–CC4 rather than CC1. SyncWords warns that not all players support reading the additional 608 channels. Check the player and destination compatibility before relying on those channels, or use a supported caption assignment for the audience you need to reach.
SyncWords’ output documentation also mentions frame reordering to support caption embedding across different frame rates and a default late-caption timeout of 30 seconds. These are advanced configuration details worth examining if the basic connection and caption settings are correct but the timing or embedding remains troublesome. They are not substitutes for checking the incoming encoding values or confirming that the intended audio is present.
For unrelated picture quality problems, such as blockiness in the upstream video, use a separate diagnosis such as the guide to fixing pixelation on a YouTube live stream. Caption configuration cannot repair a weak or malformed video source. Keep video-quality troubleshooting separate so that changes do not accidentally obscure whether the caption relay itself is working.
If you run multiple outputs or languages, keep a simple configuration record: which input source feeds the Service, which caption language is assigned to which channel, what buffer is selected, and which destination is connected. SyncWords’ billing page says usage depends on connected Event or Service time and the number of outputs, with each language’s captions, subtitles, or audio translation counted as a separate output. Confirm current contract terms directly with SyncWords before budgeting; do not assume the output count has no operational or billing consequence.
Before committing, compare the operating options on the pricing page. When the file and channel are ready, start free — 24-hour trial, no card.
FAQ
Does my encoder send directly to YouTube when I use this workflow?
No. In the documented route, your existing encoder sends its RTMP PUSH stream to SyncWords, and SyncWords sends the configured output to YouTube. Keep the input and destination URLs and keys distinct.
Do I need to enable captions in both SyncWords and YouTube Studio?
Yes. Configure transcription and Embedded 608 for the SyncWords output, then enable Closed Captions in YouTube Studio and select Embedded 608/708 as the Captions Source. SyncWords documents that the passed-through stream will not contain captions if the YouTube caption settings are not enabled as required.
Does this workflow guarantee accurate or perfectly timed captions?
No. This article explains the documented service handoff and does not claim hands-on testing, caption accuracy, or guaranteed synchronization. Rehearse with representative content and monitor the viewer-facing stream, especially for names, local terms, and music.
Can I use extra caption channels or send to other destinations?
SyncWords documents CC2–CC4 assignments and multiple RTMP destinations, but destination and player compatibility should be checked before you rely on them. This guide covers the documented YouTube workflow only; consult SyncWords’ current documentation for the configuration relevant to your account.