1. 1Download Go Live Plus from the Downloads page — macOS or Windows.
2. 2macOS: drag it into Applications, then right-click the app and choose Open the first time (it is not notarised by Apple yet). Windows: run the installer and choose More info → Run anyway if SmartScreen appears.
3. 3The app shows a short code and opens your browser.
4. 4Approve the code on the device page. The computer stays signed in until you revoke it.
5. 5The plan badge in the app should now match the plan shown here.

Installer size varies by operating system. Check the download page for size and requirements; sign-in and online voice/AI providers require internet access.

1. 1Record 15–20 seconds of yourself looking at the camera and talking normally.
2. 2Hold the head fairly still, keep the face lit evenly, and avoid hair across the mouth.
3. 3Vertical 720×1280 matches TikTok and Facebook without cropping.
4. 4Upload it on the Avatars page. Processing takes a few minutes and runs on your own machine.
5. 5When it says Ready, the avatar can be picked by any live host.

Avatar and video processing runs on your computer. Saved audio can sync to your account; text and audio go to your chosen voice/AI providers when required.

1. 1Open Live Hosts in the app and create one.
2. 2Pick the avatar you made.
3. 3Pick a voice. Paste your ElevenLabs key to use cloned voices, or use the built-in Edge voices for free.
4. 4For a voice that is not in your ElevenLabs account, use “Paste a voice ID” — a raw ID or the voice page URL both work.
5. 5Write the personality: who the host is, how it speaks, what it must never say.
6. 6Press Test to hear a line before going anywhere near a live.

1. 1Open Set up live in the sidebar. Green is done, yellow is advice, red blocks Go live.
2. 2Press Fix on any step: the fix opens inside the step — pick the host, switch products on, pick the script, choose where the live goes.
3. 3Choose the room type on the left: Avatar host, Pre-recorded, Real camera + AI voice, Q&A assistant only, or Multi-host.
4. 4Set the planned length. The script's depth is judged against it; under 60 % it sounds like a recording.
5. 5The script step shows unique minutes, one-pass minutes and whether a checkout unit exists.

Real camera + AI voice and Q&A assistant only are voice-only: your camera is the picture, the AI is heard through a hidden output link. Lipsync on a live webcam is not available yet.

1. 1Press Go live in the console or on Set up live.
2. 2Read the rows. Red blocks; yellow is advice (a short script, no rehearsal, low memory).
3. 3Rehearse first turns the same click into a rehearsal. Go live starts the live; the script starts 8 seconds later on its own.
4. 4The output link for TikTok LIVE Studio appears under Advanced as soon as the live runs; paste it once as a Browser source.

1. 1Pause script / Next unit (keys P and N) sit on the strip; Answer at the next sentence lets buying questions cut in at a sentence end.
2. 2Quick buttons queue a unit after the current one: Urgency + CTA, Re-entry hook, Checkout demo. Scheduled lines show beside them.
3. 3Take over (key T) types a line the host says next; Push to talk (Space) uses your voice. Product card switches the card.
4. 4Answered / greeted counts update live. Two minutes before the planned end the strip warns; with End on its own on, the live stops a minute after.

1. 1End live shows the report; the last ten are kept in data/reports.
2. 2Unanswered questions link to the FAQ inbox — answer them there and the host knows next time.
3. 3No sales or GMV figures: the engine cannot see orders, so it never invents them.

1. 1Press Rehearse on Set up live (or Rehearse first on the sheet).
2. 2Type a viewer comment in the box above the queue and hear the host answer; try a price question and a question it cannot know.
3. 3Press Next unit once to hear the script move on; open the Products tab and show a card.
4. 4End the rehearsal: the report appears, and Set up live shows the Rehearsal step green. Now Go live.

Demo traffic (ten canned comments) is a switch in the rehearsal box; it is off unless you turn it on.

1. 1Open Studio → Hosts → New host. Choose a distinct name, a prepared local avatar, voice, language and personality; click Save host. Repeat for the second host if needed.
2. 2Select the main host again. Enable Go live with co-hosts, click Add co-host and select the second saved host.
3. 3Set the role, Answers in languages and Owns products. Leave Voice only unchecked to show the avatar picture.
4. 4Click Save co-hosts. This is separate from Save host. If the list is empty, create and save a second host first; two library avatars do not create two hosts.

Maximum: one main host plus two co-hosts. A prerecorded-mode co-host uses voice only. Smooth dual-HD output is not guaranteed on every computer.

1. 1Click Open Multi-host Studio. Choose Aspect: 9:16, 16:9 or 1:1 before starting a session.
2. 2Select an avatar outline. Drag to move or pull the blue corner to resize. Adjust Crop, layer and Lock as needed; use Undo / Redo to correct changes.
3. 3The top-bar Background colour belongs to the whole stage. For one host’s background, select that host and use This avatar background.
4. 4Original preserves the source video. Chroma key removes the selected colour. Auto remove (local) separates the person on your computer; inspect hair, hijab, hands and products for cut-outs.
5. 5Click Save setup. The editor previews real prepared avatar frames, not moving lips; check motion in Rehearsal.

1. 1Choose Script + Q&A for scripts and viewer answers without automatic banter or follow-ups.
2. 2Choose Hybrid to allow short dialogue between script units and a one-line follow-up according to your settings. Start with infrequent dialogue; hosts take turns speaking.
3. 3In scripts, choose Who speaks for each unit or variant. Alternatively write Name: line, for example Mei: Let’s look at this product. Names must match saved hosts.
4. 4Check product facts, prices and stock before the session. A role such as Product expert divides the work; it does not automatically verify facts.
5. 5Save co-hosts, then rehearse the script with the same main host.

1. 1Under Who answers viewers’ questions, choose Auto for the default rules, or select main host, take turns, by product or by language if that suits your session.
2. 2With Auto, an explicitly named host is selected first. Next comes a recognised product and its assigned host. Recognised usage, ingredient or size questions may go to the expert; otherwise the main host answers.
3. 3In rehearsal, test a question naming a host, a question about the co-host’s assigned product, and a general question. Check both the selected speaker and its voice.
4. 4If the wrong host answers, check names, product assignments and routing settings. Selection uses rules and text matching; it does not perfectly understand every intention.

1. 1Before rehearsal, choose Aspect in Multi-host Studio and Save setup. Use 9:16 for portrait, 16:9 for landscape or 1:1 for square.
2. 2In OBS Browser Source, match width and height to the output. 720p examples: 720×1280, 1280×720 or 720×720. For 1080p, use the corresponding actual 1080p output dimensions.
3. 3Start rehearsal with the same main host. Preview, output link, recording and RTMP frames share the stage; the receiving platform can still letterbox or crop.
4. 4During an active custom-stage session, use Console → Hosts → Edit stage → Apply to live to change position, size, crop or background within the same aspect. Save setup alone does not change the running session.
5. 5To change aspect or switch between a custom stage and a preset, stop streaming and recording, end the session, then start a new one after saving the setup.

1. 1Start with two Standard-quality hosts. Check alternating lines, answers, voices, lips, avatar edges and performance. Reduce hosts or background processing if the computer is overloaded.
2. 2Sign in to the same account and organisation on web and desktop. Scripts, host metadata, voice choices, stage setups and saved audio sync; provider keys are encrypted at rest on the server.
3. 3Avatar and prerecorded video files remain local. On a second computer, import/prepare the right assets or select an available local avatar before rehearsal. A synced asset name is not the video file.
4. 4Let sync finish before switching computers. If a version conflict or audio quota issue appears, check its status rather than assuming every file is available.
5. 5Once rehearsal and platform output look right, choose Go live yourself. The AI badge remains visible.

Dual-HD performance and long sessions are not guaranteed on every computer. A visible prerecorded co-host plus avatar composite is not supported.

1. 1Open My Products and add each item with a photo, price and short description.
2. 2Fill in the keywords a customer would actually type: the brand, the nickname, common misspellings.
3. 3In the Studio, step 3 “Products & promotion”, switch on the items you are selling in this live. Max 10 — only those are in the host's knowledge.
4. 4Turn on “Place the card on the preview” to show the product card on screen. Drag it to move, pull the red corner to resize.
5. 5The host can raise a card by itself when its answer mentions that product, and you can always force one up or take it down by hand.

The AI only knows what you write here. An empty description makes it invent details, and invented prices are the fastest way to a refund.

1. 1First make the lines: sidebar → Promotion Audio → New. Give it a name, choose the host that speaks it, and type one line per row (Add line). Add pause inserts a silence of the length you set.
2. 2Set the shortest and longest pause (20 and 60 seconds by default). The host waits a different length each time, which is what stops it sounding automated.
3. 3Press Save, then Generate the audio. Wait for the badge to say Ready — a promotion that has not been generated does not appear in the Studio.
4. 4Now go to the Studio → step 3 “Products & promotion” → the “Promotion audio” box.
5. 5Choose your promotion in the dropdown (it reads “Off” until you do), then tick “Say these between answers”. That is the switch.
6. 6Go live. The host reads one line whenever it has been quiet — it never speaks over an answer.

The box only appears in Avatar Live - Lipsync mode. A prerecorded clip already has someone talking in it, so a promotion would be two voices at once.

1. 1TikTok: in the Studio, step 4 “Read comments”, paste your TikTok username and press Connect. No extension needed.
2. 2Facebook / YouTube / Instagram: install the extension from the Downloads page — unzip it, open chrome://extensions (Edge: edge://extensions), turn on Developer mode, click Load unpacked and choose the unzipped folder.
3. 3Keep the studio running on the same computer, open your live page in that browser, then pick the tab in the extension side panel and click Connect reader.
4. 4Comments land in the conversation with their platform badge, and the host answers them by priority: buying questions first, then questions, then chatter.
5. 5Turn auto-reply off at any time and the comments still arrive — the host just stops answering by itself.

The extension talks only to the studio on your own computer, over localhost. Comments are not sent to our servers.

1. 1In the Studio, press Go live, then copy the output link that appears under step 2.
2. 2In TikTok LIVE Studio, add a Browser source and paste the link.
3. 3Set the size to 1080×1920 and let it fill the canvas.
4. 4The link follows the live host, so it stays the same tomorrow and only has to be pasted once.

If it says “waiting for live stream”, nothing is live in the studio yet. Press Go live there first — the link is a window onto a running session, not a way to start one.

1. 1On the platform, create a live video and choose “streaming software”. Copy the stream key (Facebook and custom servers also give a server URL).
2. 2In the Studio, step 2 “Push to platforms”, paste the key into that platform's row and press Push.
3. 3Press Go live in the studio first, wait for the platform's own preview to show video, then press Go live on the platform itself.
4. 4“Push to every platform with a key” starts all of them at once; Stop all ends them together.

Start the stream in the studio first. Facebook ends a live that has no incoming video.

1. 1Add each destination before going live.
2. 2Press Go live once. The same video and voice go everywhere.
3. 3Comments from every platform land in the same queue.

Each platform is one account slot. A live on three platforms uses three at once — Starter covers one, Pro covers two. Live hours are still counted once.

1. 1Open Content → Pre-Records and upload your clip. Preparation starts on its own and runs in the background: Convert, Map face, Analyse speech, Mouth references, Check.
2. 2When ready, open the Lipsync editor. Green timeline ranges allow lipsync; red ranges block it. Review the automatic scan, adjust ranges as needed and save before rehearsal.
3. 3Try an answer: click a moment on the strip, type a line and press Try. You hear it in the host's voice and see the mouth in the clip say it — a rehearsal, never broadcast.
4. 4On the Hosts page set the host's Face to Pre-Recorded and pick the clip. Any voice works: Gemini, ElevenLabs, MeloTTS, Edge or OpenAI.
5. 5In the Live Console choose that host — the mode switches to Pra-rakam by itself. Rehearse first, then Go live.
6. 6While live, the Pre-Record chip shows this computer's profile, the frame rate and whether an answer is running. “Hold replies” keeps comments queued without answering, for a part of the clip you want untouched.

No “Okay, one sec” acknowledgement and no promotion audio in this mode: the clip already has a voice, and the original is muted only once per answer.

1. 1Face the camera. Your face should fill at least a quarter of the frame height; frame from the chest up.
2. 2Keep hands, the microphone and the product away from your mouth — show the product beside your face, not in front of it.
3. 3One long take: no cuts, no zooms, no change of camera angle. Turning your head a little is fine; side profiles are skipped.
4. 4Pause between sentences. An answer starts in one of your pauses, so a clip with natural breaks answers sooner.
5. 5Quiet room, or a steady music bed. Your own voice is muted while the host answers and comes back at the next pause.
6. 6Even light on the face, no strong light behind you. One person in frame.
7. 7Record 720p–1080p at 25 or 30 fps, MP4 (H.264). Phone clips (HEVC) are converted for you during preparation.
8. 820–60 minutes is enough: the clip loops smoothly. Up to 4 hours is accepted.

Below 60 % usable the Pre-Records page says what to re-record. Grey parts of the strip still play — an answer there is heard with the mouth left as it is.

1. 1Look at the live preview: if an amber “PC busy” chip is showing, this computer cannot lipsync every frame and the app is already coarsening the mouth to keep the live in the present.
2. 2Close what else is running — especially browsers with many tabs, games and video editors.
3. 3A smaller avatar frame (720×1280 rather than 1080×1920) is the single biggest saving.
4. 4Push to fewer platforms at once; each one costs an encoder.
5. 5If the chip never appears but the picture still stutters at the viewer's end, that is upload bandwidth — lower the platform's bitrate or use a cable instead of Wi-Fi.

A live that drifts further behind the longer it runs is always the machine, never the network. The app keeps itself in real time by making the mouth coarser rather than letting the delay grow.

1. 1Press Mute on the live preview in the studio.
2. 2That silences only your own speakers. The stream and the output link keep their audio.

1. 1Check Usage: if hours are at zero, the plan is spent until the period rolls over.
2. 2Check accounts in use: another live already holds the slot. End it, or move up a plan.
3. 3If the plan says expired, renew it on the billing page.
4. 4If the app says the computer is not authorised, it was revoked — sign in again with a new code.

1. 1Sign in to the same account and organisation on web and desktop. Check account connection status and let sync finish.
2. 2Scripts, host/voice settings, co-host setups, saved audio and provider keys follow your account. Keys are encrypted at rest on the server.
3. 3Avatar and prerecorded video files stay on the original computer. Import or select local assets before rehearsing on another computer.
4. 4If a sync conflict appears, do not delete local work to force a download. Check the versions on both devices first.

1. 1Restarting the app clears the build queue, which is what the top of the Avatars page shows.
2. 2Your finished avatars are in the list below it, read from disk.
3. 3If one is genuinely missing, it is still in the app's data folder and can be re-imported.

1. 1What it is
2. 2Three parts
3. 3Room to platform
4. 4Platforms

1. 1Open Downloads
2. 2Pick an installer
3. 3First-run warnings

1. 1Steps on the Downloads page
2. 2The code in the app
3. 3Computers signed in

1. 1Ready to go live panel
2. 2Read the cards
3. 3Pick the quality
4. 4HD or Standard

1. 1Live
2. 2Content
3. 3Settings

1. 1See the plans
2. 2Switch RM / USD
3. 3Compare plans
4. 4In your dashboard

1. 1Seats depend on your plan
2. 2Admin: runs the team, rooms and keys
3. 3Operator: goes live and edits content
4. 4Viewer: can only look

1. 1Only time connected to a platform counts
2. 2Preview and setup are free
3. 3Daily bars show hours per day
4. 4A 0 is a test account, not an error

1. 1Open Rooms
2. 2Type a name
3. 3Create room

1. 1Live sessions in a room
2. 2Add a live session
3. 3Open Studio

1. 1Open the defaults
2. 2Live type
3. 3Host, product, script
4. 4Save

1. 1The Room type panel
2. 2Which one?

1. 1Overview
2. 2Live sessions
3. 3Reports

1. 1Host = avatar + voice + persona
2. 2Pick an avatar that is ready
3. 3Edge is free, no key needed
4. 4Save the host, then pick it in the Console

1. 1Edge: free, no key
2. 2MeloTTS: local, Malay
3. 3Gemini, ElevenLabs, OpenAI: need a key
4. 4Try it first in the Voice Lab

1. 1Persona = instructions about the host style
2. 2The greeting is the opening line
3. 3Emotion, tone and register can be chosen
4. 4Save the host to use it

1. 1Record 20 to 30 seconds, at least 8
2. 2A quiet room, no music
3. 3Tick consent: it is your own voice
4. 4The number of clones depends on your plan

1. 1A clip of 15 seconds to 2 minutes, one face
2. 2Look at the camera, even light, mouth uncovered
3. 3Tick consent and write your full name
4. 4Building takes a while, and you can close the page

1. 1Host: pick Automatic, HD quality or Standard
2. 2HD: clearer teeth and lips
3. 3A new avatar is built Standard first, HD after
4. 4Not recommended: the live uses Standard

1. 1Preview before you pick
2. 2Three-dot menu: rename, primary, re-process, delete
3. 3Delete cannot be undone
4. 4Mouth blend mode: Blend is recommended

1. 1The idea
2. 2Open the guide
3. 3Filming tips
4. 4The target

1. 1New clip
2. 2Read the tips
3. 3Prepare
4. 4Read the result

1. 1Open a ready clip
2. 2Read the strip
3. 3Pick a range
4. 4Mark it
5. 5Save the areas

1. 1Open a ready clip
2. 2Time and host
3. 3Type a sentence
4. 4Press Try
5. 5See the result

1. 1The right name and price
2. 2The host only says what you type
3. 3Switch on In this live
4. 4Save, the card builds itself

1. 1Keywords show the card
2. 2FAQ gives fixed product answers
3. 3This image is only a sample
4. 4Save to render the card again

1. 1A script is scenes, a scene is units
2. 2One unit can hold many variants
3. 3{a|b} picks a phrase, [x|y] picks a word
4. 4Vary uses AI credit

1. 1Pick the product, language and style
2. 2Long script: 15 to 240 minutes
3. 3Tone changes how the host speaks
4. 4Real offer only, no invented urgency

1. 1Paste text or a file, it is split by line
2. 2Recordings need the HD pack or a key
3. 3Unique content meter: aim for 60% and above
4. 4Add units or variants when the meter is low

1. 1One pair: group, ways to ask, answers
2. 2Try a comment before going live
3. 3Safe mode: answer only from Q&A and prices
4. 4Match strictness: higher means stricter

1. 1AI Host pack: honest answers when asked if the host is AI
2. 2Objections pack: price, COD, shipping, size and returns
3. 3Edit the packs to match your shop policy
4. 4The inbox: Add or Ignore

1. 1Greet up to five names in one line
2. 2The per-minute cap stops the host chattering
3. 3Templates use {name}, {names}, {gift} and {count}
4. 4Press Save after changing

1. 1Scheduled line: every N minutes
2. 2Give a label and a few variants
3. 3Quick buttons show in the Console
4. 4Press Save so it takes effect

1. 1A red warning on a risky line
2. 2Categories: off-platform, giveaway, claims, competitors, false urgency
3. 3Compliance settings are in Interactions
4. 4Rules for Malaysia and Southeast Asia are not confirmed

1. 1Open Set up live
2. 2Read the progress
3. 3Press Fix
4. 4The rest

1. 1Depth meter
2. 2Planned length
3. 3Go live sheet

1. 1Pick TikTok LIVE Studio
2. 2Paste the link

1. 1Pick a platform
2. 2Key in Advanced
3. 3Saved keys

1. 1The Virtual camera option
2. 2How it works

1. 1The AI badge step
2. 2Locked on
3. 3Platform rules

1. 1Open Platforms
2. 2Saved stream keys
3. 3Add or change

1. 1Check Rehearsal
2. 2Start Rehearsal
3. 3Wait for the avatar
4. 4Type a question
5. 5End live

1. 1The Now strip
2. 2Pause script
3. 3Next unit
4. 4Take over
5. 5Say exactly this
6. 6Cut

1. 1Comment queue
2. 2Demo traffic
3. 3Priorities
4. 4Answer rules
5. 5Buttons on a row
6. 6The Auto switch

1. 1Script tab
2. 2Switch script
3. 3Products tab
4. 4Place the product card
5. 5Setup tab

1. 1Open the sheet
2. 2Read the eight rows
3. 3Choose an action

1. 1Strip chips
2. 2Memory warning
3. 3Preview badge
4. 4Check the system

1. 1P: pause the script
2. 2N: next line
3. 3T: take over text
4. 4Space: press to talk
5. 5E: end live, with a confirmation

1. 1TikTok LIVE comments
2. 2Username
3. 3Rehearsal with demo traffic
4. 4Auto answer

1. 1Open Advanced
2. 2Auto answer
3. 3Twitch
4. 4Facebook and Instagram

1. 1What there is
2. 2Video over RTMP
3. 3Comments
4. 4What you can do

1. 1Scripts
2. 2Pick a language
3. 3Without an AI key
4. 4Scene and unit
5. 5Q&A
6. 6Interactions

1. 1Media triggers
2. 2New rule
3. 3Placement
4. 4Try a sentence
5. 5Test in a rehearsal

1. 1Open the Canvas
2. 2Layers
3. 3Add text
4. 4Drag and resize
5. 5Lock, hide, reorder
6. 6Save

1. 1Scene bar
2. 2New scene
3. 3Switch scenes
4. 4Template gallery
5. 5My templates

1. 1Media library
2. 2Upload files
3. 3Folders and tags
4. 4Trim a clip
5. 5Replace a file

1. 1Safe-area guide
2. 2TikTok and Shopee
3. 3Stream to
4. 4Compliant portrait
5. 5AI host label

1. 1Live layers
2. 2Viewer question
3. 3Thanks
4. 4Show live output
5. 5From the console

1. 1New host
2. 2Avatar and voice
3. 3Save host
4. 4Co-hosts
5. 5Save co-hosts

1. 1Open the editor
2. 2Pick a format
3. 3Drag and resize
4. 4Crop
5. 5Save setup

1. 1Conversation mode
2. 2Who answers
3. 3Other host adds
4. 4Host banter
5. 5Save

1. 1Room type
2. 2Pick Multi-host
3. 3Read the check
4. 4Back to Avatar host

1. 1Sample comments
2. 2Type a question
3. 3End live
4. 4Read the report

1. 1Filter 7, 30 or 90 days
2. 2Open a row for the session report
3. 3Only what Studio can measure
4. 4Export CSV for your own spreadsheet

1. 1After a live, check Questions to answer
2. 2Write the real answer, then press Add
3. 3That question is now answered instantly
4. 4Repeat after every live

1. 1What it is
2. 2Events in
3. 3Host events out
4. 4Custom RTMP
5. 5Two routes

1. 1Address and token
2. 2Start a rehearsal
3. 3List sessions
4. 4Send a comment
5. 5Limits

1. 1What MCP is
2. 2Connect
3. 335 tools
4. 4AI prepares, you press
5. 5Limits

1. 1Look at the health strip
2. 2The PC busy chip
3. 3Fix it

1. 1Find the sound button
2. 2Mute the preview

1. 1Open the Go live sheet
2. 2Read the Memory row
3. 3Fix it

1. 1Check Account
2. 2Four usual reasons

1. 1The same account
2. 2Check the lists
3. 3What follows you

1. 1Open Advanced
2. 2The TikTok LIVE comments card
3. 3Read the status