Put in a photo and a recording, and out comes a video with the lips matched up—LipsyncX compresses this into three steps. In its own tutorial, it records a calculation: a fictional art teacher’s photo, an 8-second course-opening recording, 480p selected, two models each generating once, 20 credits deducted each time. The 8 seconds are counted as two 5-second blocks, because the billing rule is “each started 5 seconds is billed as a full block.” Source[1] This bill was checked on October 14, 2026. The rates on the current pricing page (retrieved 2026-11-14) are: 2 Fast 10 credits/5 seconds, 2 Pro 480p 20 credits/5 seconds. The two 20s in the tutorial were the actual deductions at the time; at the current Pro rate, the same 8 seconds would be deducted 40 today. [2]
That is the delivery in the tutorial: getting a face that cannot move to match 8 seconds of sound.
Image: source LipsyncX actual image source[3], retrieved 2026-11-10, original interface preserved.
The official site compresses usage into three steps: upload a clear photo of a person’s face or an existing video; enter a script, upload audio, or record directly in the browser; click generate, check the result, and the finished video can be reopened from creations. Source[4] The site also gives a specific use case: take an existing video and replace the soundtrack; a translation workflow can export a localized version with subtitles; the site says it supports lip-sync and video dubbing in more than 50 languages. [4]
What the Product Takes Over Is the “Match the Lips” Step
It does not say it handles everything. A troubleshooting article on the site puts it plainly: a successful render does not mean it is ready to publish; people still need to check the visuals and sound; the article also does a calculation based on the smallest pack at $2.99/60 credits—that balance covers three attempts of this kind, but not three videos guaranteed to be publishable. [5]
What the product takes over is combining a photo and a recording into a video with matching lips; after export, a person still has to see whether the lip movements and sound fit, and then decide whether to publish. Someone who, as in the tutorial, uses a ready-made photo and a recording to make a course opening saves the mechanical labor of repeatedly matching lip movements; what is not saved is choosing the material and taking a final look.
Candor on the Price List
Its charging method reveals a set of trade-offs: someone who makes an occasional video and someone willing to subscribe monthly get different terms.
A one-time pack of 60 credits sells for $2.99, no subscription required, and the credits do not expire; any successful top-up, including this starter pack, unlocks watermark-free downloads of existing and future videos. Source[2] New accounts get 20 trial credits on login, and trial videos carry a watermark. Source[4] The trial flow lets you make a watermarked finished version first and then decide whether to pay to unlock it; conversely, buying credits first and generating afterward also works.
The other tier is Pro Monthly, $39 a month, with 1,350 base credits per paid month, automatically renewing until cancellation; unused subscription credits carry over for only one billing cycle and expire on their original expiry date. Source[2] On the same site, the one-time pack “does not expire,” while the subscription pack “carries over for only one period”—two sets of rules written side by side on the pricing page.
Billing granularity is 5 seconds, and each started 5-second block is added as a full block; the higher the resolution and the more expensive the model, the more credits are deducted. The page shows the total price before generation. Source[2] When rendering fails on the provider side, deducted credits are returned to the account balance—a credit return, not a cash refund. [1]
What is rare is the underlying-model column. Under LipSyncX 2 Pro, the pricing page states that the underlying model is InfiniteTalk (MeiGen-AI, based on Alibaba Wan 2.1, via WaveSpeedAI), and makes clear that the product name identifies only the service tier and workflow, not a claim to have developed the underlying model in-house. Source[2] Putting the supplier’s name on the price list is equivalent to saying the product-tier name does not mean an in-house underlying model; what users buy is the orchestration and packaging layer. This candor is worth acknowledging, and it also conveniently marks the competitive position the product admits to.
A Small Ledger and a For-Sale List
TrustMRR’s revenue table verified with the Stripe API (last synced 2026-11-07) lists $908 for the last 30 days and $8,642 all time. Source[6] The Metric Snapshots section on the same page lists current MRR $130, MRR growth of -28.3% over the last 30 days, and active subscriptions 6. Source[6] Six active subscriptions are not the total number of customers; they only indicate the size of the subscription line; Startup Insights separately lists estimated user count 6,300, also an estimate.
current MRR $130 and total revenue of $908 over the past 30 days are two different measures; TrustMRR lists the revenue table and Metric Snapshots separately, and the share from one-time credits is not broken out. The last 30 days’ profit margin of 80.0% is also in Metric Snapshots. The page does not give a cost-accounting method, so it is impossible to judge whether model invocation fees, maintenance, and payment-channel fees are included; 80.0% is only a TrustMRR metric snapshot and cannot be treated as profit independently verified through Stripe.
The business is now on the shelf. TrustMRR’s Acquisition section says: Listed for sale, first listed 2026-10-27, asking price $18,000, revenue multiple 1.6x, and 2 offers received. Source[6] Profile lists country HK and founded 2024-03-22; Startup Insights lists team size 1 person. Source[6] Marketing channels are listed as blog, content marketing, SEO—the two earlier articles, the tutorial and the troubleshooting article, are products of this channel setup. [6]
Alternatives are not scarce. HeyGen’s official talking-photo page describes the same thing as: a single image can generate a voice-synced video (Avatar IV), claiming coverage of 175+ languages and dialects. Source[7] The shared condition is: both turn one still photo into a talking video. LipsyncX also has a starter pack at $2.99/60 credits, billed in 5-second blocks, with credits returned on failure; HeyGen’s page does not give a price, so it cannot be said here which is more expensive. Another player being present does not mean every user should choose it; the choice can rest on 175+ languages and dialects, or on a three-dollar start and credits returned for provider-side failures.
A site that sells credits writes an article telling you not to waste credits on re-renders; a product that lists its supplier names on the price sheet has now also put itself on a for-sale list. The asking price on TrustMRR is $18,000—that is the asking price for the entire project, not the price of any one service tier, and not yet a closing price.
[1] How to Choose an AI Talking Video Model for Your First Clip
[2] LipsyncX Pricing | Pay-As-You-Go AI Lip Sync Video
[3] LipsyncX actual image source
[4] Lip Sync AI Video Generator Online — Free Trial
[5] AI Lip Sync Looks Wrong? Fix the Input Before You Render Again